Aberystwyth University
Streamer Waves Dissertation Project
Presented for BSc Astrophysics
Mr Nicholas David Mellor Institute of Mathematics, Physics and Computer Science
Nicholas David Mellor Streamer Waves 130053235
1
Abstract
A program was devised and written using IDL, to calculate the properties of the oscillation of
a coronal streamer after a shock front from a coronal mass ejection (CME) collided with it. Two CME/
streamer collision events were examined, on the dates 23/06/2012 from 15:48 to 16:24 and
21/04/2013 from 07:12 to 11:00. After the limit of the electron density data was reached, the event
on 23/06/2012 from 15:48 to 16:24, had to be discarded. However, from the remaining event, results
were gathered. These results displayed a positive relationship for the wave speed and Alfvén speed
with increasing heliocentric height, and a negative relationship for ion plasma density, magnetic field
strength and magnetic tension force with increasing heliocentric height.
Nicholas David Mellor Streamer Waves 130053235
2
Table of Contents Abstract 1
Table of Figures 3
1. Introduction 4
1.1. Experimental Purpose 4
1.2. Introduction to the Sun, the Solar Corona, Coronal Streamers and Coronal Mass Ejections 4
1.3. Introduction to SoHO, LASCO-C2, IDL and SolarSoftWare 6
2. Literature Review 7
2.1. Introduction 7
2.2. Magneto-Hydrodynamics (MHD), Coronal and Alfvén Waves 7
2.3. CMEs and Magnetic Reconnection 11
2.4. Streamers, the Slow and Fast Solar Wind and LASCO 12
2.5. Conclusions 13
2.6. Bibliography 13
3. Assumptions and Equations 15
3.1. Assumptions made about the Streamer Model 15
3.2. Assumptions made about the Calculations 15
3.3. Equations used in the Experiment 15
4. Experimental Method 17
4.1. Data Acquisition 17
4.2. Data Manipulation and Calculations 18
4.2.1. Data Manipulation 18
4.2.2. Calculations 19
4.3. Data Analysis 22
5. Results, Discussion and Error Analysis 23
5.1. Results 23
5.2. Discussion 25
5.3. Error Analysis 27
6. Conclusion and Acknowledgements 30
6.1. Conclusion 30
6.2. Acknowledgements 30
7. Additional Literature 31
7.1. Bibliography 31
8. Appendices 33
8.1. Appendix 1 33
8.2. Appendix 2 34
Nicholas David Mellor Streamer Waves 130053235
3
8.3. Appendix 3 41
Table of Figures Figure 1: Four different MHD oscillation modes of cylindrical fluxtubes: fast MHD sausage mode (a),
fast MHD kink mode (b), slow MHD (acoustic) mode (c) and MHD torsional mode (d). (Aschwanden,
2006) 10
Figure 2: Diagram of a chord of a circle. 16
Figure 3: Two images, one from 2012/06/23 showing an example of a 'thin' streamer on the lower
right of the image; another image from 2013/10/31 showing an example of a 'fat' streamer on the
lower left of the image. 17
Figure 4: An example of original file used in the experiment. 18
Figure 5: An example of a manipulated file used in the experiment. 18
Figure 6: A flowchart to show the process the IDL program follows. 19
Figure 7: An example of a running difference image (RDI) used in the experiment 20
Figure 8: A diagram to show points along the wavelength 21
Figure 9: A graph showing wave speed against height from solar centre. 23
Figure 10: A graph to show Alfven speed against height from solar centre. 23
Figure 11: A graph to show ion plasma density against height from solar centre. 24
Figure 12: A graph to show magnetic field strength against height from solar centre. 24
Figure 13: A graph showing magnetic tension force against height from the solar centre. 25
Figure 14: A diagram showing wave propagation. 25
Figure 15: An image showing the oscillation along the streamer on 21/04/2013 from 07:12 to 11:00.
26
Figure 16: An image showing the oscillation along the streamer on 23/06/2012 from 15:48 to 16:24.
26
Figure 17: A table comparing calculated and literature wave speeds. 27
Figure 18: A table to show further wave speed results. 27
Figure 19: A graph showing the electron density along the normalised heliocentric height, at an angle
of 578o of the data structure. 28
Figure 20: A graph to show the average ion plasma density with heliocentric height. 28
Figure 21: A table comparing the ion densities to the average ion densities. 29
Figure 22: A table displaying all the data gathered during the experiment. 33
Figure 23: A table showing some average results. 33
Figure 24: A graph showing the wave frequency against heliocentric height. 33
Nicholas David Mellor Streamer Waves 130053235
4
1. Introduction
1.1. Experimental Purpose
The purpose of this experiment was to determine the physical properties of the oscillation of
coronal streamers (from this point referred to simply as streamers), after a collision with a shock wave
from a coronal mass ejection (CME). The observed properties include the wavelength and wave speed
of the propagating wave. The properties from the calculations; include the Alfvén speed, the magnetic
field and the magnetic tension force along the wave, the frequency of the fast kink mode, and the time
period of the wave. These latter properties were determined using coronal electron plasma density
values, provided by Dr Huw Morgan at Aberystwyth University (Quémerais and Lamy, 2002; Morgan,
2015). Two dates were considered and measured during this experiment, 23/06/2012 from 15:48 to
16:24 and 21/04/2013 from 07:12 to 11:00. These dates and times were used as both contained CMEs
occurring close to a streamer, this gave a strong and clear oscillation of the streamer as the shock wave
collided with it. The times taken were from as the CME starts, to when the streamer stopped
oscillating.
The secondary aim for this experiment is to provide another method of measuring the
streamer waves’ physical properties. Very little literature can be found solely about calculating these
properties, as such the aim was to provide a method of calculating approximate values for a simple
fundamental streamer wave. This method can then be built upon to give a more complete and accurate
picture of the coronal streamer. As such, an IDL program (henceforth known as “the program”) was
written to carry out all of the calculations needed to return the required results.
1.2. Introduction to the Sun, the Solar Corona, Coronal Streamers and Coronal Mass
Ejections At approximately 4.5 billion years old and of type G2V (Guenther, 1989), the Sun is the centre
of our solar system. It is a population I star, or a metal-rich star; all elements above hydrogen and
helium in the periodic table are considered metals, in astronomical terms. The sun has a metallicity of
Z=0.0166, this suggests that the Sun could have formed from a supernova explosion, where the metals
found in the Sun would have formed. There is no definite boundary to the Sun, in its outer regions the
density decreases exponentially as the distance from the centre increases (Zirker, 2002). However, for
measurement purposes the Sun’s radius is taken to be the distance from its centre to the edge of the
photosphere, which is the apparent visible surface of the Sun (Phillips, 1995).The Sun goes through
solar cycles of varying solar activity, beginning approximately every 11 years. A solar minimum is the
time period of the least number of sunspots and solar flares, often neither will be spotted for days at
a time during this period. A solar maximum is the opposite. It is the time period with the greatest solar
activity (Moussas, Polygiannakis, Preka-Papadema and Exarhos, 2005).
The solar corona is an aura of plasma surrounding the sun that extends millions of kilometres
into space. During a total solar eclipse, the corona is most apparent this is due to the moon blocking
out all light from the solar disk, as the corona is far dimmer than the disk. As such, instruments used
to observe the solar corona utilize an artificial eclipse to make the corona more visible. The corona is
a region of very hot plasma, much hotter than the surface of the sun, with a maximum temperature of
approximately 1 million Kelvin (Markus, 2004). The light in the corona is produced by 3 primary
sources; The K-corona, which is created when sunlight is scattered by free electrons in the corona. The
F-corona, when sunlight is reflected and scattered by dust particles, observable because the light
contains Fraunhofer absorption lines. Fraunhofer lines are dark absorption lines in the Sun’s spectrum
Nicholas David Mellor Streamer Waves 130053235
5
that can be detected when the full visible spectrum is analysed. And the E-corona, which is created by
the spectral emission lines from the ions that are present in the plasma (Ridpath, 2012). For this
experiment, it has been assumed that the only light being received is light due to the K-corona, this is
due to the Doppler broadening of the absorption lines completely obscuring them, creating the
appearance of a continuous spectrum with no absorption lines (Golub and Pasachoff, 2010). The solar
corona’s structure is in three-parts, comprising of: dark coronal holes, where the fast solar wind
originates as plasma escaping outward along open magnetic fields. Bright coronal loops that connect
regions in the photosphere that are of opposite polarity, these are magnetically closed. Finally, smaller,
intense regions called X-ray bright points, that consist of tiny loops and are scattered over the solar
disk (Priest, 2014).
Coronal streamers are the most prominent feature in the solar corona. Due to the high energy
electrons trapped within their magnetic field lines, they stand out against their background and have
a very definitive boundary. Streamers often tend to overlie active regions or prominences and are an
equilibrium between the confining magnetic field and the expanding plasma. They consist of a base
(or an arcade) of closed field lines, that is topped by a blade or a fan of open field lines. Streamers are
formed when the upper reaches of a large coronal arcade extend high enough to be stretched open
by the solar wind. Helmet streamers can exist in two types of magnetic topology (Hundhausen, 2012)
(Zhao and Webb, 2003). A unipolar streamer or pseudo-streamer, lies over a quadrupolar field;
meaning two oppositely directed bipolar fields adjacent to each other (Wang, Biersteker, Sheeley Jr et
al., 2007) (Hundhausen, 2012), and separates coronal holes of the same polarity. Or bipolar streamers
that lies above a singular bipolar arcade and between coronal holes that have opposite polarities,
creating an overlying blade in the form of a current sheet with oppositely directed magnetic fields on
both sides. During the solar minimum, streamers tend to appear from the mid-latitude regions. As the
minimum starts to turn to the maximum the streamers begin migrating to the poles, and the slow solar
wind is then seen from all latitudes (Chen, 2013).
Streamers are thought to follow the magnetohydrodynamic (MHD) kink mode (Chen, Song, Li
et al., 2010), as seen in figure 1. This means that the magnetic field is arranged in a cylinder that is
almost incompressible, and can cause a displacement of the axis of the plasma structure, or more
simply the streamer will be displaced from its equilibrium position. MHD modes guided by the tube,
with phase speeds larger than the minima of the Alvfén and sound speeds ( 𝑣𝐴𝑖 , 𝑣𝐴𝑒 , 𝑐𝑠𝑖 , 𝑐𝑠𝑒) are called
fast, whereas those smaller than one or the other minimum are called slow. The sound speed is
dependent on temperature and the Alfvén speed depends on the density and magnetic field. For fast
kink waves in a tube, the phase speed is the kink speed, 𝑐𝐾 = [𝐵𝑖2 + 𝐵𝑒
2]/[𝜇(𝜌𝑖 + 𝜌𝑒)]1
2 , which is the
mean Alfvén speed (Nakariakov and Verwichte, 2005). The frequency of this mode is given by the
equation 𝜔𝐾 = √2𝑘𝑧
2𝐵2
𝜇(𝜌𝑖+𝜌𝑒) (Nakariakov, Melnikov and Reznikova, 2003; Doorsselaere, Nakariakov and
Verwichte, 2008). It is this displacement that is being investigated, as such, it is assumed that streamers
follow the fast kink mode model.
Coronal mass ejections (CMEs) are sudden and often violent releases of solar material into
interplanetary space. There are two types, fast and slow CMEs. Fast CMEs release energy in a more
explosive manner than slow CMEs, however they do not show signs of accelerating unlike slow CMEs
(Gosling, Bame, McComas et al., 1994; J. T. Gosling, 1996). Coronal mass ejections have an associated
shock wave preceding them, it is this shock wave that interacts with the streamer.
Nicholas David Mellor Streamer Waves 130053235
6
1.3. Introduction to SoHO, LASCO-C2, IDL and SolarSoftWare The Solar and Heliospheric Observatory (SoHO) was launched on 2nd December 1995. Its
purpose; to study the internal structure of the Sun, its outer atmosphere and the origins of the solar
wind. SoHO is a collaboration between the European Space Agency (ESA) and the National Aeronautics
and Space Administration (NASA) (Domingo, Fleck and Poland, 1995). It was originally planned only to
be a two-year mission but it is still operational over 20 years later. SoHO is in a halo orbit about the
Sun-Earth L1 point. This is the point between the Sun and the Earth, where the balance of the Earth’s
gravity and the Sun’s gravity, equals the centripetal force required for an object to have the same
orbital period as the Earth around the Sun. This results in the object staying in that relative position
(Domingo, Fleck et al., 1995). A consequence of this, is that the Earth does not impede SoHO’s ability
to capture images of the sun, resulting in it being able to monitor the Sun constantly.
One of the instruments aboard SoHO is the Large Angle and Spectrometric Coronagraph
(LASCO) this consists of three coronagraphs:
C1- a Fabry-Pérot interferometer coronagraph, which images from 1.1-3 solar radii.
C2- a white light coronagraph imaging from 2-6 solar radii, which uses an orange filter.
C3- a white light coronagraph imaging from 3.7-32 solar radii, that uses a blue filter.
The solar corona is monitored by these coronagraphs by implementing an optical system to create an
artificial eclipse, to block out the light from the solar disk. Both the C2 and C3 coronagraphs produce
images over most of the visible spectrum, whereas the C1 interferometer produces images of the
corona in very narrow visible wavelength bands (G. E. Brueckner, 1995). For the purposes of this
experiment, only the C2 coronagraph was used, as the other two were either of the wrong distance
range, or wrong spectrum or both.
Interactive Data Language (or IDL) is a programming language used for data analysis and image
manipulation. It is popular in various areas of science such as medical imaging, atmospheric physics
and astronomy. IDL has been used in many space projects and missions, for example the calibration of
high-energy X-ray telescopes and the rescue of the Hubble Space Telescope. IDL’s use in Earth sciences
include being used in improving the accuracy of long-range weather predictions, and the tracking of
sea ice to keep ships safe (Stern, 2000). There are many libraries of IDL procedures, all of them contain
procedures that are specific to a certain area of science. There are however, a few astronomical
libraries containing similar procedures but maintained by different organisations. Examples include;
the IDL astronomy user’s library by the Goddard Space Flight Center (Center), the Coyote library by
Coyote’s guide to IDL programming (Consulting) and SolarSoftWare by S.L. Freeland and R.D. Bentley
(Freeland). IDL originated from an early version of Fortran, this can be seen through the similar use of
syntax and the base integer being 0 and not 1. IDL is owned and maintained by Exelis Visual Information
Solutions, Inc., a subsidiary of Harris Corporation (Corporation).
SolarSoftWare (SSW) is a set of integrated software libraries, system utilities and databases
which provide a common programming and data analysis environment for solar physics. SSW is
primarily an IDL based system, it is a collection of data management and analysis routines obtained
from the SoHO and the Yohkoh missions, the Solar Data Analysis Center and the Astronomy libraries,
along with other packages (Freeland and Handy, 1998). All of the SSW routines can then be called into
the IDL program once the environment has been loaded. In this experiment, IDL was used with the
SSW environment installed, this allowed for the required data manipulation to be carried out.
Nicholas David Mellor Streamer Waves 130053235
7
2. Literature Review
2.1. Introduction Streamers are a very prominent structure in the Solar Corona, and understanding the physical
properties of these streamers can be vital to the understanding of the functions and dynamics of the
corona and sun itself. The effect of the magnetic field of the streamers is an interesting phenomena in
which the streamers appear to oscillate during coronal mass ejection events. The goal of this project
is to calculate various physical properties of steamer waves, such as; phase speed, the magnetic
restoring force and the period of oscillation. This literature review was carried out in preparation for
the project in order to better understand the various properties and phenomena of the corona and
the sun.
2.2. Magneto-Hydrodynamics (MHD), Coronal and Alfvén Waves Magneto-hydrodynamics (MHD) is the study of the magnetic properties of electrically conducting
fluids, such as plasmas. The basic concept of MHD is that; magnetics fields in a moving conductive fluid
can induce currents within that fluid. These currents polarizes the fluid, which in turn, changes the
magnetic field itself.
Energy is carried upward through the solar interior by sound waves generated in the hydrogen
convection zone, this is below the photosphere. These waves propagate into the “fast” mode,
becoming increasingly magneto-hydrodynamic in nature as they pass through the photosphere, out
into the chromosphere. This is due to the negative density gradient. Very little energy is emitted by
the hydrogen convection zone in the “slow” or “Alfvén” modes as these modes are strongly absorbed
by the photosphere. Cross-sections for collisions between ions and neutral atoms in the chromosphere
is large, meaning that the dissipation of the fast-mode waves by the frictional damping mechanism is
very small. These waves build into shock waves, the dissipation of which provides the main energy
source for the chromosphere (Osterbrock, 1961). Energy, in the form of acoustic energy, reaches the
corona preferably along almost vertical magnetic fields, these vertical fields are inside the vertical flux-
tubes. Most of the magnetic energy stays concentrated beneath the transition region due to the
refraction of the magnetic waves, as well as the continuous conversion of acoustic-like waves into fast
magnetic waves; at the equipartition layer, located in the photosphere where plasma beta equals 1.
However, some of the magnetic energy (that propagates in the region where the arcades are located)
does reach the low corona (Withbroe and Noyes, 1977) (Santamaria, Khomenko and Collados, 2015).
Plasma beta, β, is the ratio of the plasma pressure to the magnetic pressure, this can be expressed by
the following equation (K. L. DeRose, 2008):
𝛽 =𝑝
𝑝𝑚𝑎𝑔=
𝑛𝑘𝐵𝑇
𝐵2(2𝜇0)⁄
The coronal magnetic field has multiple effects on the hydrodynamics of the coronal plasma. It can
play a passive role, meaning the magnetic geometry does not change. For example, the magnetic field
can maintain thermal insulation between the plasmas of adjacent loops or flux-tubes, or it can channel
plasma and heat flows or particles and waves along the magnetic field lines. However, the magnetic
field can also play an active role, such that the magnetic geometry does change, for example; the build-
up and storage of non-potential energy, the triggering of an instability, the changing of topology by
the means of various types of magnetic reconnection, or by exerting a Lorentz force on the plasma.
The acceleration of plasma structures such as, filaments, prominences and coronal mass ejections
(CMEs) are also a result of the magnetic field playing an active role.
Nicholas David Mellor Streamer Waves 130053235
8
MDH is expressed in terms of macroscopic parameters, such as pressure, density, temperature, and
flow speed of the plasma. In the macroscopic scale the plasma reacts to electric and magnetic forces
as described by the Maxwell equations. The particle motion of the plasma, however, can also be
described by microscopic physics, called Kinetic theory, in terms of the Boltzmann equation or the
Vlasov equation(Aschwanden, 2005).
Many astrophysical plasmas are characterized by the ideal MHD equations, which includes the MHD
continuity equation (1), the momentum equation (2), Maxell’s equations (3-5), Ohm’s law (6) and a
specialized equation of state for energy conservation (7). For example, incompressible, isothermal, or
adiabatic. The full set of ideal MHD equations are shown below, for an adiabatic equation of
state(Aschwanden, 2005):
1. dρ
dt= −ρ∇. 𝐯
2. ρd𝐯
dt= −∇p − ρg + (𝐣 x 𝐁)
3. d
dt(pρ−γ) = 0
4. ∇ x 𝐁 = 4π𝐣
5. ∇ x 𝐄 = −1
c
∂𝐁
∂t
6. ∇. 𝐁 = 0
7. 𝐄 = −1
c(𝐯 x 𝐁)
A magnetic flux tube is defined as a surface, generated when a set of field lines intersect a simple
closed curve. The magnetic flux that crosses a section S of the flux tube is given as:
𝐹 = ∫𝑩.𝑑𝑺
𝑆
EQ.1
From the divergent free nature of the magnetic field and using Gauss’s law, the following can be
obtained:
∫∇.𝐁
V
dV = −
−F1
∫ 𝐁d𝐬
S1
⏞ +
F2
∫ 𝐁d𝐬
S2
⏞ + ∫0
𝐁. �̂�⏞ dσ
tubesurface
= 0 EQ.2
Where the flux tube surface is parallel to the magnetic field lines, so, 𝐁. �̂� = 0. As such,∇. 𝐁 = 0.
Therefore, it can be can be stated that F1=F2, this means that the magnetic flux tubes conserves the
magnetic flux. However, at the flux tube surface, when embedded in a field-free or weak-field medium
a discontinuity occurs that can be described as the following step function:
𝜃(𝑥) = {1 𝑤ℎ𝑒𝑛 𝑥 > 00 𝑤ℎ𝑒𝑛 𝑥 < 0
It can be calculated that 𝐣∗ =1
μ[B], from this it can be shown that a current sheet occurs at the flux-
tube surface, where the current flows perpendicular to the magnetic force lines. Then using the
momentum equation and Ampère’s law;
ρd𝐯
dt= −∇p + (𝐣x𝐁) + ρ𝐠
∇x𝐁 = μ𝐣
Nicholas David Mellor Streamer Waves 130053235
9
It is possible to obtain:
Where “i” and “e” refer to the interior and exterior of the flux-tube, but in the immediate vicinity of
the surface. This result shows a discontinuity in pressure occurs at the flux-tube surface, such that the
Lorentz force is balanced by another pressure force (Roberts and Webb, 1978). This equation also
shows the total pressure, consisting of gas pressure plus magnetic pressure, is constant across the flux-
tube’s surface. (Steiner, 2006)
The following is the procedure for calculating the magnetic structure of a vertical, axisymmetric and
untwisted magnetic flux-tube. It is worth noting however, that this is calculated for a perfectly
symmetric object, which probably does not occur in reality. These steps can still be followed for more
complex situations.
i. An initial magnetic configuration must be specified. This flux-tube will be embedded within a
plane-parallel atmosphere which is thought to remain unchanged.
ii. Calculate gas pressure in all space using hydrostatics along magnetic force lines, 𝑝 =
𝑝0(Ψ) exp [−∫𝑑𝑧′
𝐻(𝑇(Ψ,𝑧′))
Ψ,𝑧
Ψ,0]. Since this does not solve an energy equation, the pressure scale
height as a function of field line and height, 𝐻(Ψ, 𝑧), must also be specified. It is acceptable to
guess that H is independent of Ψ and identical to scale height of the surrounding field-free
atmosphere, this is approximately filled if the temperature is constant in the planes parallel to
the solar surface. From the initial, embedding atmosphere the pressure and pressure scale
height in the surrounding field-free atmosphere can be found.
iii. Evaluate the current sheet and volume using 𝑗𝜙∗ =
2(𝑝𝑒−𝑝𝑖)
𝐵𝑖+𝐵𝑒 and 𝑗𝜙 = 𝑟
𝜕𝑝
𝜕Ψ|𝑧.
The first equation applies to the locations of discontinuous gas pressure and field strength,
such as the surface of the flux-tube. The second applies to regions of continuous gas pressure.
iv. Solve the Grad-Shafranov equation, 𝜕2Ψ
𝜕𝑟2−1
𝑟
𝜕Ψ
𝜕𝑟+𝜕2Ψ
𝜕𝑧2= −𝜇𝑟𝑗𝜙, using appropriate boundary
conditions. This will give a new field configuration, Ψ(r, z), that cannot necessarily be equal to
the initial magnetic configuration. Therefore, steps (ii) to (iv) must be reiterated until
convergence has been achieved (Steiner, 2006).
There are four distinct kinds of MHD modes, all of which have different propagation, dispersion and
polarisation properties (Nakariakov, Melnikov et al., 2003) (Doorsselaere, Nakariakov et al., 2008).
These different modes are listed below:
Sausage modes, these are oblique fast magnetoacoustic waves guided by a plasma structure.
This mode causes the contraction and the expansion of the plasma structure, but the axis of
which is not displaced. These modes can be compressed and can cause significant variation of
the modulus of the oscillating structures magnetic field. For sausage modes, the azimuthal
wave number, m, equals 0. This would be shown as a tube with fixed end points, with the
space in between expanding and contracting. The frequency of this mode is given as:
𝜔𝑆 = √𝑘𝑧2𝐵2
𝜇𝜌𝑒
Kink (or transverse) modes, these are also oblique, fast magnetoacoustic waves guided by a
plasma structure. Unlike sausage modes, kink modes are weakly compressible and do cause
the displacement of the axis of the plasma structure. They could be imaged by instruments as
𝑝𝑒 +
𝐵𝑒2
2𝜇= 𝑝𝑖 +
𝐵𝑖2
2𝜇 EQ.3
Nicholas David Mellor Streamer Waves 130053235
10
periodic propagating or standing displacements of coronal structures. The frequency of this
mode is given as: 𝜔𝐾 = √2𝑘𝑧
2𝐵2
𝜇(𝜌𝑖+𝜌𝑒)
Torsional (Alfvén or twist) modes are transverse, incompressible perturbations of the magnetic
field along certain individual magnetic surfaces. Unlike kink modes, torsional modes cannot be
observed with imaging instruments, this is because they do not affect the displacement of
either the plasma structure axis or its boundary. The frequency of torsional modes is given by:
𝜔𝐴 = √𝑘𝑧2𝐵2
𝜇𝜌𝑖
Longitudinal (slow or acoustic) modes, are slow magnetoacoustic waves which propagate
mainly along the magnetic field lines of the plasma structure. These modes are compressible
and the magnetic field perturbation in these modes is negligible. The frequency of this is given
below, where CS is the sound speed, and CA is the Alfvén velocity: 𝜔𝐿 = √𝑘𝑧2 (
𝐶𝑆2𝐶𝐴
2
𝐶𝑆2+𝐶𝐴
2)
Below is figure 1; it shows each of the above mentioned modes as wire cylinders. It shows in detail
how the plasma structure of each mode can change.
Figure 1: Four different MHD oscillation modes of cylindrical fluxtubes: fast MHD sausage mode (a), fast MHD kink mode (b), slow MHD (acoustic) mode (c) and MHD torsional mode (d). (Aschwanden, 2006)
An Alfvén wave in a plasma is defined as a low frequency travelling oscillation of the ions and the
magnetic field. The wave propagates with the magnetic field lines; waves do exist at oblique incidence,
the interface between dielectric media, and smoothly change into a magnetosonic wave when
propagation is perpendicular to the magnetic field. The ion mass density provides the inertia, and the
tension of the magnetic field lines provides the restoring force. The wave is also dispersionless, and
the motion of the ions and the perturbation of the magnetic field, are in both the same direction, and
transverse to the direction of propagation. Alfvén waves are also considered as suitable transporters
of non-thermal energy, which is required to heat the Sun’s inactive atmosphere (Hasegawa and Uberoi,
1982) (Mathioudakis, Jess and Erdélyi, 2013).
The Alfvén velocity can be found, by knowing the low-frequency relative permittivity, ε, of a
magnetised plasma. This can be calculated by:
𝜖 = 1 +1
𝐵2𝑐2𝜇0𝜌
Nicholas David Mellor Streamer Waves 130053235
11
Where, c is the speed of light, µ0 is the permeability of free space, B is the magnetic field strength
and 𝜌 = ∑𝑛𝑠𝑚𝑠 is the total mass density of all the charged plasma particles; s denotes all plasma
species, including electrons and most ions. From this, the phase velocity of an electromagnetic wave
in this medium can be calculated:
𝑣 =𝑐
√𝜖=
𝑐
√1 +1𝐵2𝑐2𝜇0𝜌
Or, 𝑣 =
𝑣𝐴
√1 +1𝑐2𝑣𝐴2
[EQ. 4]
The Alfvén velocity is given by;
𝑣𝐴 =𝐵
√𝜇0𝜌 [EQ. 5]
2.3. CMEs and Magnetic Reconnection Coronal Mass Ejections (or CMEs) are the sudden releases of solar material into interplanetary space.
There are two classified types of CME, fast and slow (Gosling, Bame et al., 1994; J. T. Gosling, 1996).
Fast CMEs show no sign of acceleration, with the energy released in a more explosive manner. Slow
CMEs which accelerate in the direction of the chronograph suggests a continuous ejection of matter
and energy. The average energy released in a CME is between 1022 and1024𝐽(Howard, Sheeley,
Koomen and Michels, 1985). Based on SoHO and LASCO observations between 1996 and 2003 the
speed of the CMEs have been measured to vary greatly, from 20 to 3000km/s, and their transit time
to Earth’s radius can range from a few hours to a few days (V. Yurchyshyn) (V. Bothmer, 1996). CMEs
directed towards Earth are known as Halo CMEs due to their circular ‘halo’ appearance on the
chronograph. Full halo events extend 360O around the sun. A partial halo event on the other hand, is
one with an apparent width that is greater than 120O, and appears in projection above at least one
pole (Hudson, Lemen, St. Cyr et al., 1998; Burlaga, Plunkett and St. Cyr, 2002). The frequency of CMEs
are very dependent on the phase of the solar cycle, during solar minimum about one CME every 5 days
is expected, this increased to between 3 and 4 every day.
CMEs appear to originate predominantly from the mid latitude regions and from active regions such
as sunspots. CMEs were originally thought to be driven by the heat of an explosive flare, however it is
now suggested that they are unrelated. This is because the energy released in a CME must logically be
built up over time and released quickly, the mechanism of this release is still up for debate, but one
idea is magnetic reconnection. CMEs produce evident shock waves that disrupt local and global solar
structures, this occurs when the speed of the CME exceeds the local Alfvén speed (Raymond,
Thompson, St. Cyr et al., 2000).
The majority of CMEs, 60-70%, appear to emerge from flux ropes embedded in coronal streamers and
seem to occur when large amounts of plasma collect in the streamer tips (Subramanian, Dere, Rich
and Howard, 1999) (Chen, 2013). Many CMEs appear to display a flux rope topology characteristic, an
explanations for this is, on occasion, some field lines embedded deep within the CME seem to only
have one end connected to the sun (Gosling, Birn and Hesse, 1995).
Magnetic reconnection seems to play a vital role in the explosive events in the corona, especially in
solar flares. A reconnection is driven by two oppositely directed inflows which collide in an area known
Nicholas David Mellor Streamer Waves 130053235
12
as the diffusion region. In this diffusion region the plasma β is high but the magnetic field is low. (Priest,
Forbes and Forbes, 2000) It is still unclear however how the energies produced play into CMEs. On the
reverse of this, CMEs may drive reconnections along streamer current sheets.
2.4. Streamers, the Slow and Fast Solar Wind and LASCO Streamers are the most prominent feature in the global coronal structure, they are clear against their
background with a definitive brightness boundary. They are much brighter than the background
corona; this is due to the high energy electrons trapped in magnetic loops. Other properties of
streamers include closed magnetic field arcades, streamer cusps, and high density plasma sheets(Chen,
Song et al., 2010). Streamers are only visible during a solar eclipse, or by the use of a chronograph.
They tend to overlie active regions and are a balance between the confining of the magnetic field and
the expanding plasma.
Streamers are also thought to be a source for the slow solar wind and during solar minimum (period
of lowest solar activity) tend to appear from the mid latitude regions. As the minimum wanes,
streamers migrate towards the poles and the slow solar wind is then seen from all latitudes. (Chen,
2013) The slow solar wind averages 400km/s at 1AU and as mentioned appears to come from
streamers, some evidence for this source of solar wind is through magnetic reconnection in streamers,
giving to a release of matter and energy. Specifically this is outside the streamer-coronal hole
boundaries where the velocity of the wind depends greatly on the magnetic topology of flux tubes. In
contrast however, the fast solar wind is thought to originate in coronal hole regions and averages
speeds between 700 and 800km/s at 1AU. (U. Feldman, 2005).
Streamers have been thoroughly observed through the use of LASCO-C2 observations. Past
observations indicated that the magnetic structure of streamers consist of closed magnetic arcades,
however more recent data from LASCO-C1 seem to suggest a series of closed loops
The first part of a streamer is considered as a bulge or a helmet, and they typically stretch to around
1.5 solar radii. The width and radial extension of streamers is also dependent on solar cycle, being
smaller and shorter around solar maximum, the converse is true for solar minimum (Koutchmy and
Livshits, 1992). The second noticeable structure in a streamer, is the stalk. The stalk can stretch over
many solar radii, and is an effect of both the solar wind and the magnetic properties within the
streamer.
Large Angle Spectrometric Chronograph (or LASCO), is the primary instrument used to observe the
solar corona. It is based on The Solar and Heliospheric Observatory (or SoHO), and is often used in
tandem with another instrument on-board SoHO, the Extreme ultraviolet Imaging Telescope (or EIT).
LASCO consists of 3 different instruments, known as C1, C2 and C3. The LASCO-C1 is a Febry-Pérot
interferometer which images from 1.1 to 3 solar radii. C2 and C3 are both white light chronographs
which image from 2 to 6, and 3.7 to 32 solar radii respectively (G. E. Brueckner, 1995). The Ultraviolet
Coronagraph Spectrometer on the SoHO satellite has detected coronal emission lines of H, N, O, Mg,
Al, Si, S, Ar, Ca, Fe and Ni, especially in coronal streamers. The instrument covers the 940-1350Å range
and the 470-630Å range in second order(Raymond, Kohl, Noci et al., 1997).
CMEs and streamers have an intrinsic relation to one another, whilst streamers are thought to be a
potential source for a CME, CMEs give rise to large-scale phenomena on nearby streamers. During a
CME nearby streamers can be deflected and give rise to a streamer wave; a long-period, large scale
wave, which carries energy and propagates outwards along the plasma sheet. The damping of this
wave has nothing to do with dissipative processes(Spruit, 1982), it is instead due to the effect of the
Nicholas David Mellor Streamer Waves 130053235
13
magnetic field lines. The well-known Lorentz equation can provide a link between fluidic equations and
electromagnetic;
𝑗̅ × �̅� =
1
𝜇(𝐵. ∇)𝐵 − ∇(
𝐵2
2𝜇), EQ.3
Where μ is the magnetic permeability This can give rise to a magnetic restoring force, the effect of which means the amplitude of the wave
decreases with time. This can also be known as the magnetic tension;
Tension Force =
1
𝜇
𝐵02
𝑎2(𝑥𝑖 + 𝑖) EQ.4
Some measurements of the magnetic field of streamers show that it decreases from 0.3G at 3 solar
radii, to 0.01G at 10 solar radii(Chen, 2013). The magnetic properties of the streamer can be deduced
if the phase speed of the wave is known, given various parameters such as the period. EQ. 2 can be
used to calculate the phase speed of an electromagnetic wave.
2.5. Conclusions Over the course of this literature review, the processes of various wave properties within the Sun’s
interior and atmosphere have been looked at, as well the origins of coronal streamers and the
mechanisms which cause wave oscillations within the streamers. It has been determined that CMEs
drive streamer waves, as well as streamers being a potential source for CMEs. Also the damping of
these waves is due to the magnetic field lines within streamer itself, rather than dissipative forces.
Furthermore, it was discovered that using the LASCO C2 and C3 instruments on board SoHO are the
best way to view streamers.
2.6. Bibliography 1. Osterbrock, D.E., The Heating of the Solar Chromosphere, Plages, and Corona by
Magnetohydrodynamic Waves. The Astrophysical Journal, 1961. 134: p. 347. 2. Withbroe, G.L. and R.W. Noyes, Mass and Energy Flow in the Solar Chromosphere and Corona.
Annual Review of Astronomy and Astrophysics, 1977. 15(1): p. 363-387. 3. Santamaria, I.C., E. Khomenko, and M. Collados, Magnetohydrodynamic wave propagation
from the subphotosphere to the corona in an arcade-shaped magnetic field with a null point. Astronomy and Astrophysics, 2015. 577: p. 70.
4. K. L. DeRose, C.C., P. Pribyl, W. Gekelman. Measurement of the Plasma Beta in the Enormous Toroidal Plasma Device (ETPD). 2008; Available from: http://www.pa.ucla.edu/sites/default/files/files/REU/Papers%202008/derose_kim.pdf.
5. Aschwanden, M.J., Physics of the Solar Corona. An Introduction with Problems and Solutions (2nd edition). Pour la Science, ed. M.J. Aschwanden. 2005. 853.
6. Roberts, B. and A.R. Webb, Vertical motions in an intense magnetic flux tube. Solar Physics, 1978. 56(1): p. 5-35.
7. Steiner, O., Photospheric Processes and Magnetic Flux Tubes. 2006. 8. Nakariakov, V.M., V.F. Melnikov, and V.E. Reznikova, Global sausage modes of coronal loops.
Astronomy and Astrophysics, 2003. 412: p. L7-L10. 9. Doorsselaere, T.V., V.M. Nakariakov, and E. Verwichte, Detection of Waves in the Solar Corona:
Kink or Alfvén? The Astrophysical Journal Letters, 2008. 676(1): p. L73. 10. Aschwanden, M.J., Coronal magnetohydrodynamic waves and oscillations: observations and
quests. Philosophical Transactions of the Royal Society of London A: Mathematical, Physical and Engineering Sciences, 2006. 364(1839): p. 417-432.
Nicholas David Mellor Streamer Waves 130053235
14
11. Hasegawa, A. and C. Uberoi, Alfven wave. DOE Critical Review Series. 1982. p. Medium: ED; Size: Pages: 137.
12. Mathioudakis, M., D.B. Jess, and R. Erdélyi, Alfvén Waves in the Solar Atmosphere. From Theory to Observations. Space Science Reviews, 2013. 175: p. 1-27.
13. J. T. Gosling, P.R., The acceleration of slow coronal mass ejections in the high-speed solar wind. Geophysical Research Letters, 1996. 23(21).
14. Gosling, J.T., et al., The speeds of coronal mass ejections in the solar wind at mid heliographic latitudes: Ulysses. Geophysical Research Letters, 1994. 21(12): p. 1109-1112.
15. Howard, R.A., et al., Coronal mass ejections: 1979–1981. Journal of Geophysical Research: Space Physics, 1985. 90(A9): p. 8173-8191.
16. V. Yurchyshyn, S.Y., H. Wang, N. Gopalswamy, Statistical Distributions of Speeds of Coronal Mass Ejections. Big Bear Observatory, Centre for Solar Physics ans Space Weather, Crimean Astrophysical Observatory, Centre for Solar Research of New Jersey Institute of Technology, Laboratory for Experimental Physics: Califonia, USA. Washington, USA. Crimea, Ukraine. New Jersey, USA. Maryland, USA.
17. V. Bothmer, R.S., Signatures of Fast CMEs in Interplanetary Space. Advances in Space Research, 1996. 17(4).
18. Burlaga, L.F., S.P. Plunkett, and O.C. St. Cyr, Successive CMEs and complex ejecta. Journal of Geophysical Research: Space Physics, 2002. 107(A10): p. SSH 1-1-SSH 1-12.
19. Hudson, H.S., et al., X-ray coronal changes during Halo CMEs. Geophysical Research Letters, 1998. 25(14): p. 2481-2484.
20. Raymond, J.C., et al., SOHO and radio observations of a CME shock wave. Geophysical Research Letters, 2000. 27(10): p. 1439-1442.
21. Subramanian, P., et al., The relationship of coronal mass ejections to streamers. Journal of Geophysical Research: Space Physics, 1999. 104(A10): p. 22321-22330.
22. Chen, Y., A review of recent studies on coronal dynamics: Streamers, coronal mass ejections, and their interactions. Chinese Science Bulletin, 2013. 58(14): p. 1599-1624.
23. Gosling, J.T., J. Birn, and M. Hesse, Three-dimensional magnetic reconnection and the magnetic topology of coronal mass ejection events. Geophysical Research Letters, 1995. 22(8): p. 869-872.
24. Priest, E., T. Forbes, and T. Forbes, Magnetic Reconnection. Magnetic Reconnection, by Eric Priest and Terry Forbes, pp. 612. ISBN 0521481791. Cambridge, UK: Cambridge University Press, June 2000., ed. E. Priest. 2000.
25. Chen, Y., et al., Streamer Waves Driven by Coronal Mass Ejections. The Astrophysical Journal, 2010. 714(1): p. 644.
26. U. Feldman, E.L., N. A. Schwadron, On the sources of the fast and slow solar wind. Journal of Geophysical Research: Space Physics, 2005. 110.
27. Koutchmy, S. and M. Livshits, Coronal Streamers. Space Science Reviews, 1992. 61: p. 393-417. 28. G. E. Brueckner, e.a., The Large Angle Spectroscopic Chronagraph (LASCO). Solar Physics, 1995.
162(1). 29. Raymond, J.C., et al., Composition of Coronal Streamers from the SOHO Ultraviolet
Coronagraph Spectrometer. Solar Physics, 1997. 175(2): p. 645-665. 30. Spruit, H.C., Propagation speeds and acoustic damping of waves in magnetic flux tubes. Solar
Physics, 1982. 75(1-2): p. 3-17.
Nicholas David Mellor Streamer Waves 130053235
15
3. Assumptions and Equations Throughout the project assumptions had to be made, either to make the equations or the
model simpler, and are listed below.
3.1. Assumptions made about the Streamer Model These assumptions were made about the model of the streamer, how it was being visualised,
and used.
The oscillation of the streamer is a fundamental fast kink mode, this reduces the complexity
of the model and allows for simpler calculations.
The streamer is cylindrical with a radius equal to the radius visible in the image.
Within the streamer the electron density is equal to the ion density, as the streamer is not
electrically charged.
The ion plasma only consists of hydrogen ions.
3.2. Assumptions made about the Calculations These assumptions were made during the calculations that were used throughout the
experiment.
At the distance SoHO is from the sun, it can be assumed that 1 arc-second is equal to 730km.
This allows a distance to be calculated from an image, given that the number of arc-seconds
per pixel is known.
The plasma density outside of the streamer is much less than the plasma density inside the
streamer, 𝜌𝑒𝑥𝑡𝑒𝑟𝑛𝑎𝑙 ≪ 𝜌𝑖𝑛𝑡𝑒𝑟𝑛𝑎𝑙.
The magnetic field inside the streamer and the magnetic field outside the streamer are equal,
𝐵𝑖 = 𝐵𝑒.
The wave shape follows that of a sine wave, as such the wave is symmetrical and hence the
total wavelength is double that of a half wavelength.
3.3. Equations used in the Experiment Throughout the project multiple equations are used. This section contains a list of equations
and the variables used in them, as well as some derivations.
Alfvén velocity, 𝑣𝐴 =𝐵0
√𝜇0𝜌, where, 𝐵0 is the magnetic field, 𝜇0 is the permeability of free space,
𝜌 is the plasma density. For this experiment, this equation was used in the form 𝐵0 = 𝑣𝐴√𝜇0𝜌,
to calculate magnetic field.
Kink speed, 𝑐𝐾 = (𝐵𝑖2+𝐵𝑒
2
𝜇0(𝜌𝑖+𝜌𝑒))
1
2, where the magnetic field inside the flux tube is 𝐵𝑖, and outside
the tube is 𝐵𝑒. Using the above assumptions, results in:
𝑐𝐾 = (𝐵𝑖2 + 𝐵𝑒
2
𝜇0(𝜌𝑖 + 𝜌𝑒))
12
𝑐𝐾 = (2𝐵2
𝜇0𝜌)
12
𝑐𝐾 =√2𝐵
√𝜇0𝜌= √2𝑣𝐴
So,
Nicholas David Mellor Streamer Waves 130053235
16
𝑣𝐴 =𝑐𝐾
√2
Wavenumber, 𝑘 =2𝜋
𝜆, where λ is the wavelength in meters.
The frequency of the fast kink mode (Doorsselaere, Nakariakov et al., 2008), 𝜔2 = 𝜔𝑘2 = 𝑣𝑘
2𝑘2,
where the kink speed, 𝑣𝑘 = 𝑣𝐴 [2
1+𝜌𝑒𝜌𝑖
]
1
2
, 𝜌𝑒 is the electron density, 𝜌𝑖 is the ion density and k is
the wave number. Substituting known values gives:
𝜔2 = 𝑘2𝑣𝐴2 [
2
1 +𝜌𝑒𝜌𝑖
]
𝜔 = √2𝑘2𝑣𝐴
2
1 +𝜌𝑒𝜌𝑖
𝜔 = √2𝑘2𝐵0
2
𝜌𝑖𝜇0 (1 +𝜌𝑒𝜌𝑖)
And hence
𝜔 = √2𝑘2𝐵0
2
𝜇0(𝜌𝑖 + 𝜌𝑒)
Time period, 𝑇 =2𝜋
𝜔.
Lorentz force, 𝒋𝑥𝑩 =(𝑩.𝛁)𝑩
𝜇− 𝛁(
B2
2𝜇), the first term on the right- hand side is the magnetic
tension force, the second term on the right-hand side is the magnetic pressure force. Only the
magnetic tension force is needed in this experiment.
Magnetic tension force, (𝑩.𝛁)𝑩
𝜇, this is non-zero if B varies along the direction of B, and can be
regarded as being produced by the effect of a tension along B of magnitude 𝐵2𝜇⁄ per unit
area. It can also be said that 𝑩 = 𝐵�̂�, in terms of the unit vector along the field, as such:
𝑚𝑎𝑔𝑛𝑒𝑡𝑖𝑐 𝑡𝑒𝑛𝑠𝑖𝑜𝑛 𝑓𝑜𝑟𝑐𝑒 (𝑇𝑚) =𝐵
𝜇
𝑑
𝑑𝑠(𝐵�̂�)
𝑇𝑚 =𝐵
𝜇
𝑑𝐵
𝑑𝑠�̂� +
𝐵2
𝜇
𝑑�̂�
𝑑𝑠
𝑇𝑚 =𝑑
𝑑𝑠(𝐵2
2𝜇) �̂� +
𝐵2
𝜇
�̂�
𝑅𝑐
Where, �̂�, is the unit vector principal normal to the magnetic field line, and Rc, is its radius of
curvature (Priest, 2014).
Radius of curvature, 𝑅𝑐 =ℎ
2+𝑤2
8ℎ, where, h, is the height of the
circle segment, and, w, is the width of the circle segment, shown
in figure 2.
Figure 2: Diagram of a chord of a circle.
Nicholas David Mellor Streamer Waves 130053235
17
4. Experimental Method 4.1. Data Acquisition
Collecting the data for this experiment proved to be challenging. Due to a solar minimum in
2008-2009, no useful data could be collected for a couple years either side of the solar minimum
peak, from approximately mid-2006 through to mid-2010. Also, because of the nature of the
experiment, and certain parameters that must be met; such as the need to see a clear oscillation and
perturbation along the streamer, and the difficulties encountered when there is no single
predominate streamer (as shown in figure 3); the number of events that were useable were reduced
drastically. This number again was reduced when the events that were considered had no associated
images, or there was a server error and the images could not be returned.
Firstly, a NASA database of CME events and LASCO daily movies was scoured to find suitable
events (NASA). This website contains daily movies from the LASCO instrument aboard the SoHO
spacecraft, the database has information on nearly all of the CME events since 1996, information such
as acceleration, mass and kinetic energy of the coronal mass ejections. When looking for events to use,
it was the events whose mass was greater than 1x1016g. The logic behind this was that the CMEs that
contained the most mass, had the most kinetic energy and therefore had the greatest chance to
displace the streamer. Once events had been found, the associated data files then had to be obtained.
These files were downloaded in the form of FITS files from the LASCO images query form, found on the
Large Angle and Spectrometric Coronagraph Experiment website ((NRL); (NRL)). A Flexible Image
Transport System (or FITS) is an open standard that defines a digital file format useful for transmission,
storage and processing of scientific images. FITS is the most common type of digital file format used in
astronomy; as the file type is designed specifically for scientific data and each file contains a header.
This is unlike many image formats as this header includes many provisions for describing photometric
and spatial calibration information (Wells, Greisen and Harten, 1981). The term image is loosely
applied, when taken in context to FITS files; as the format supports data arrays of arbitrary dimensions,
whereas, normal images are usually 2- or 3-Dimensional. However, for simplicity in this paper, FITS
files will be split into two parts, the header information and the ‘image’ (Pence, Chiappetti, Page et al.,
2010).
Figure 3: Two images, one from 2012/06/23 showing an example of a 'thin' streamer on the lower right of the image; another image from 2013/10/31 showing an example of a 'fat' streamer on the lower left of the image.
Nicholas David Mellor Streamer Waves 130053235
18
As the brightness of the images was not used during the experiment, uncalibrated FITS images
could be used; therefore, level 0.5 C2 files were downloaded. During the download process, it had to
be made sure that all the files were of the same size, that is 1024x1024 pixels. This is due to these files
being the largest files available, and as such the resolution of these images is the greatest. Images that
have the starting resolution of 512x512 pixels, have different properties to the 1024x1024 images,
such as a different number of arc seconds per pixel and different resolutions. Because of this, the
512x512 pixel images had to be discounted, and only the 1024x1024 images could be used, due to this
the time differences between each image could be inconsistent.
4.2. Data Manipulation and Calculations
4.2.1. Data Manipulation Data manipulation is the process by
which data is changed in an effort to make it
easier to read and use. The manipulation of the
data in this experiment was relativity simple.
Starting with an original FITS file, an example of
which is given in figure 4, a black mask is added
to the image so it covers the bright ring at the
centre of the image. This was done to remove
the brightest part of the image, so that the
second part of the data manipulation can be
completed. After the mask was added, the top
and bottom 1 percent of the intensity
distribution were removed. This can be achieved
by thinking of all the intensity values, from light
to dark, being put into a histogram; the quantity
of each intensity value was then calculated and
plotted on the graph. If for arguments sake,
that the darkest intensity value has a value of
1, and the most intense value is equal to 100,
then all the values of intensity 1 and 100 are
ignored and removed. An example of the
resulting file is shown figure 5. This did not
affect the ultimate outcome of the
experiment, as stated above the brightness
was not used, it simply made the coronal
features such as streamers and coronal mass
ejections more visible.
All the data manipulation, calculations
and analysis was carried out in an IDL program,
which is provided in the appendices under
appendix 2. This code was particularly required
to carry out the calculations that this
experiment demanded; these calculations are
outlined below.
Figure 4: An example of original file used in the experiment.
Figure 5: An example of a manipulated file used in the experiment.
Nicholas David Mellor Streamer Waves 130053235
19
4.2.2. Calculations Due to the data being used in the experiment, existing in FITS file format, the basis for most of
the calculations is somewhat subjective. To tell the program which pixel coordinates to take for various
calculations, such as; the streamer length, the wavelength and radius of curvature; the user is required
to use the cursor to click the image, on the points that the program should use. These points are also
the basis for the speed calculations and as such the calculations of the magnetic field, frequency and
time period. Although the program informs the user where they should click, the exact location will
vary depending on the image, the user, and what the user perceives is the correct position of the point.
However, despite the subjective nature of the experiment in that sense, the program does
return suitable and somewhat accurate results, but this shall be discussed later. The calculations
themselves are well tested and are laid out in a logical manner. The following figure displays a flow
chart that summerises the process taken by the program, the actual code of which is given in appendix
2.
Figure 6: A flowchart to show the process the IDL program follows.
As shown in the flowchart, the first calculation required is to calculate the length of the
streamer. This was achieved by the user selecting the ‘base’ of the streamer, the region of the streamer
closest to the mask. The user then needs to select the last visible part of the streamer, the IDL program
then calculated the difference in these points in pixels. From the header of the FITS file, the program
extracts information about the number of arc-seconds per pixel, in this case 11.9, and so the number
of pixels is multiplied by this. Using the assumption that at the distance the SoHO spacecraft is from
the Sun, there is approximately 730km per arc-second, the streamer length in arc-seconds is then
multiplied by 730km. Finally, 1 solar radius is added to this value, this then gives the approximate
length of the streamer. This is the same technique that is used in future distance measurements.
The program creates a series of running difference images (RDIs) so that the kink speed can be
calculated. An RDI is the resulting image where one image is subtracted from another, for example,
RDI=image2-image1, this is shown in figure 7 as an example. The bright regions representing where
the wave is in the second image, the dark regions showing where the wave was initially. Using RDIs it
is possible to see where the streamer was and is, in two consecutive images. This means that by
Data manipulation.
Streamer length calculation.
Running difference
images created.
Kink speed calculation.
Mid-distance calculation.
Alfvén speed calculation.
Approximate wavelength calculation.
Ion plasma density
calculation.
Magnetic field strength
calculation.
Radius of curvature
calculation.
Magnetic tension force calculation.
Frequency calculation.
Time period.
Plotting the results in graphs.
Printing results to console.
Printing results to a file.
Nicholas David Mellor Streamer Waves 130053235
20
selecting two points on an RDI, e.g. the two
crests of the wave at different times, and
dividing by the time period between the
images (speed =distance
time), the speed can be
calculated. Each time the kink speed is
calculated; the average distance between the
two points from the centre of the Sun
(henceforth called mid-distance), is also
calculated and stored in an array. At each
mid-distance point, the wave/kink speed,
Alfvén speed, magnetic field strength,
magnetic tension force, ion plasma density,
time period and frequency are calculated.
Using the relation between kink speed, cK, and
Alfvén speed, 𝑣𝐴 =𝑐𝐾
√2, the Alfvén speed is
calculated; this coupled with the ion plasma
density and the permeability of free space
gives the magnetic field strength, by 𝐵0 =
𝑣𝐴√𝜇0𝜌𝑖.
The ion plasma density is calculated from the data provided by Dr Huw Morgan at Aberystwyth
University (Quémerais and Lamy, 2002; Morgan, 2015). By using the assumption that ion and electron
plasma densities are equal, the electron plasma density needs to be converted from cm-3 to m-3
(multiplied by 1x106) and then multiplied by the mass of a proton. The data structure in which the
electron density is stored is set out in a radial system, where there are 720 radial components, so data
for every 0.5 degree. Then, for each radial component, from 2 solar radii (Rʘ) to 5 Rʘ, there are 100
data points at intervals of 1 and starting at 0. As such, to select a data point, both a radial component
and a height above 2 Rʘ, is required. It is here that the mid-distance is also used, for each mid-distance
height, the electron density is selected and stored in an array to be used in the magnetic field strength
calculation. However, to comply with the layout of the data structure, the mid-distances had to be
normalised, such that the mid-distance is given in terms of integer values of 100. This was
accomplished by the following method:
x =middistance
(5Rʘ − 2Rʘ
100 )=middistance
26112.825km
Due to the maximum height constraint implemented during the creation of the electron density data
structure, any height above 5 Rʘ will not return a density value and will cause the program to crash. As
such a limit had to be put into place, so any mid-distance value above 99 would be given a value of 0,
hence there being some 0 values at higher heights. Since the ion plasma density has been calculated
at various heights above the 2 Rʘ, the magnetic field strength is also calculated for the same mid-
distance heights. This shows how the magnetic field strength within the streamer changes along the
length of the streamer.
It was also necessary to calculate the radius of curvature; this was needed in the magnetic
tension force equation. The radius of curvature calculates the radius of the circle that is contained
within the curve of the displaced streamer. The equation, 𝑅𝑐 =ℎ
2+𝑤2
8ℎ was used to calculate this, as
shown in figure 2, w, is the width of the chord, and, h, is the height of the chord. This was calculated
Figure 7: An example of a running difference image (RDI) used in the experiment
Nicholas David Mellor Streamer Waves 130053235
21
for the images where the streamer is oscillating below 5 Rʘ. Again these values were calculated using
user selected clicks, and distances calculated as described above. The calculated radius of curvature
and the magnetic field strength, B, can then be combined to calculate the magnetic tension force, Tm,
the SI units of which are Pam-1, or Pascal per meter. Using the equation mentioned in section 3.3. for
magnetic tension force, as given below, where μ is the permeability of free space:
𝑇𝑚 =𝑑
𝑑𝑠(𝐵2
2𝜇) �̂� +
𝐵2
𝜇
�̂�
𝑅𝑐
When both of the unit vectors, �̂� and �̂�, are taken to be equal 1, the equation reduces to:
𝑇𝑚 =𝐵2
2𝜇+𝐵2
𝜇
1
𝑅𝑐
Since the magnetic field is calculated at various points along the streamer, and the radius of curvature
is also calculated as the wave propagates along the streamer. It is also possible to calculate the
magnetic tension force, along the streamer.
The wavelength of the streamer is also calculated and
again the user click input method is implemented. Figure 8 is a
diagram that is showing a simplified version of the displaced
streamer, the black circles and diamonds show two sets of points
along the streamer. If the user selects one set of points the
program multiplies that calculated distance by 2, which gives the total approximate wavelength of the
streamer. From this wavelength, in meters, the wavenumber can also be calculated. In the case of this
experiment the angular wavenumber is found, using the equation 𝑘 =2𝜋
𝜆. The wave number is
required as it is needed to calculate the frequency, and subsequently the time period of the wave.
The equation for the frequency of the fast kink mode was derived in section 3.3., in the
simplest form it is given as 𝜔 = √2𝑘2𝐵0
2
𝜇0(𝜌𝑖+𝜌𝑒). Where, k is the wavenumber, ρi is the ion plasma density,
ρe is the electron density. Again, due to the magnetic field strength and the wavenumber having been
calculated at various points along the streamer, it is possible for the frequency to be calculated at the
same points along the streamer. This allows for the user to see how the frequency varies with
heliocentric height along the streamer. From the frequency of the kink mode, it is then possible to
calculate the time period of the wave, using 𝑇 =2𝜋
𝜔. This equation will return the time period in
seconds, for the purpose of the experiment it was decided to change the time period to hours. This
was due to the time intervals between each image taken by LASCO on-board SoHO, being 10. 6̇ minutes
and 12.8 minutes, as such it was assumed that the time period will be an hour or more.
Figure 8: A diagram to show points along the wavelength
Nicholas David Mellor Streamer Waves 130053235
22
4.3. Data Analysis After the program carries out all of the required calculations, it then plots the results and saves
all of the data calculated into a file. The program does both of these procedures, so that not only is the
data presented so that it is easier to see trends, it is also possible to have the exact numbers for
comparison purposes. There is more data that the program does not print to file, but instead, will print
to the console. This is due to the data either not being relevant for comparison, or because the data is
in the form of a number, that is needed for a subsequent calculation. However, it is interesting to see
the data nonetheless, hence it is printed to the console.
Unfortunately, due to the limitations of the electron density data only extending out to 5Rʘ,
and the streamer on the 23/06/2012 from 15:48 to 16:24, starting to oscillate from approximately 3Rʘ;
it was not possible to calculate results from this date. As such only the streamer on 21/04/2013 from
07:12 to 11:00 was actually used in the calculations, as this data was within the 5Rʘ limit and so would
return results.
Nicholas David Mellor Streamer Waves 130053235
23
5. Results, Discussion and Error Analysis
5.1. Results In this section the important results from the experiment are presented in graphical form. All
other results from this experiment are given in appendix 1, in the form of tables of data from all of the
graphs, as well as some data that is not in graphs.
Figure 9, shows the relationship
between the wave speed, or the kink speed,
of the propagating wave along the streamer,
and the radial heliocentric distance. As can
be seen from the figure there is, on average,
a positive relationship between these two
parameters. The points at around 3.5 and 4.5
solar radii (Rʘ), seem to be erroneously high
and low, respectively. This is mostly caused
by human error due to the subjective nature
of this program. As such these will slightly
affect the results later on.
As in the previous figure, figure 10
shows the relationship between the Alfvén
speed, against the radial heliocentric
distance. Due to the relation between the
kink speed and the Alfvén, as mentioned in
section 4.2.2. figure 9 has a similar shape to
figure 10. As such, the points that were
erroneous in figure 9 are also erroneous in
figure 10.
Figure 9: A graph showing wave speed against height from solar centre.
Figure 10: A graph to show Alfven speed against height from solar centre.
Nicholas David Mellor Streamer Waves 130053235
24
Figure 11, displays the ion plasma
density at the same radial heliocentric
distances as in previous figures. As can be
seen from the figure, there is a negative
relationship between the two parameters,
this is to be expected as the particle density
decreases with distance from the sun.
However, it can also be seen that the graph
goes to zero, this is due to the density data
structure limitations, so any height value
above 5Rʘ will return a zero-value.
In figure 12, the magnetic field
strength is shown against the radial
heliocentric height; it can be seen from the
graph, that there is a negative relationship
between the two. Because the magnetic field
strength relies upon the plasma density, the
magnetic field strength above 5Rʘ cannot be
calculated and so a zero value is given. The
first point on the graph at approximately
3.1Rʘ, is erroneously low, however, this could
be due to an exceptionally low Alfvén speed
or ion plasma density. Aside from these two
points, the graph shows a negative trend
which is as expected as the magnetic field will
decrease with increasing height.
Figure 11: A graph to show ion plasma density against height from solar centre.
Figure 12: A graph to show magnetic field strength against height from solar centre.
Nicholas David Mellor Streamer Waves 130053235
25
Figure 13 displays the magnetic
tension force against the radial heliocentric
distance. As before with figure 12, the graph
goes to zero, and the first point at
approximately 3.1Rʘ is lower than expected.
Both of these points are due to the vaules of
the magnetic field strength, because of the
equation for magnetic tension force:
𝑇𝑚 =𝐵2
2𝜇+𝐵2
𝜇
1
𝑅𝑐
This has caused the first point that was
already low, to become even more
pronounced, and the last point to be zero.
There is however, a negative relationship
that can be seen; this as expected, if the
magnetic field strength is weakening, then
less force will be required to straighten the
magnetic field lines.
5.2. Discussion As a starting point for the method used in this experiment, a paper in the Astrophysical Journal
(Chen, Song et al., 2010), was heavily used for inspiration. From an attempted replication of this earlier
study, the initial method was created. However, this replicated method was eventually abandoned,
and a new method adopted. This newer method, the one described in this experiment, arose due to
the software that was available during the experiment.
The results presented in the paper are an example of the results that are produced by the
program. These will obviously change with regards to the streamer/ CME shock collision event, and
the electron plasma density. However, these results all
show roughly the correct changes, based on the
calculations, as they move from higher to higher distances.
There are some obvious anomalies, these were mentioned
in the results section as they were presented. These
anomalies can be avoided, but for the sake of the experiment, they were included to give some insight
as to why they would arise, and how they can be avoided.
For example, when selecting the points to use for the speed calculations, it is important to
select the peak of the wave, as indicated in figure 14 by the black circles, as the peak propagates along
the streamer. As mentioned in the calculations section, this selection was carried out using running
difference images; as such, the wave can be in what is effectively two places at two different times in
one image, as seen in figure 7. The difficulty with this method is due to the image subtraction. If the
same pixel in the two images contains the same information, then that information is lost when the
subtraction takes place. This means that it can be hard sometimes to judge where the peak is, and so
the user has to decide. This is where the subjective nature of the experiment arises. This is also why
the wave speed varies so much and why anomalies occur, if the speed equation is examined, speed =distance
time, if the user judges the two peaks to be further apart than they actually are, then the speed will
Figure 13: A graph showing magnetic tension force against height from the solar centre.
Figure 14: A diagram showing wave propagation.
Nicholas David Mellor Streamer Waves 130053235
26
increase and vice versa. However, another reason that the wave speed varies, is a problem that cannot
be avoided easily. This is the amount of time between each image from LASCO. In the case of the date
focussed upon in this experiment, there were two different time periods between images. This has
already been touched on, these are 10. 6̇ minutes and 12.8 minutes. Depending over which of these
time periods the two images where taken, will also affect the overall wave speed. These two reasons
would explain the anomalous points at approximately 3.5Rʘ and 4.5Rʘ in figure 9, and subsequently in
figure 10.
As previously mentioned, the basis of
gathering data points from the FITS files in this
experiment, is purely subjective. It is up to the
user to judge where they should click on the
image, when prompted to do so. Whilst this does
return results, the results are not always
consistent; this could be due to the program not
registering a click, or the program accepting one
click for two sets of points, for example. In this
sense the program is not very robust. An
improvement to the program would be
implementing a piece of edge detection
software. Exelis Visual, the company behind IDL,
have some imaging software called ENVI
(Corporation), that can carry out edge detection
procedures, unfortunately this software was not
available to use during this experiment.
For this experiment to be even more
successful, a greater range of electron density
data would be required. If the electron density
data was available out to 10Rʘ for example, it
would then be possible to study the propagation
of waves along the streamer, as in figure 16,
using both the LASCO-C2 and LASCO-C3
instruments. This would provide a greater and a
more reliable understanding of streamers and
magnetohydrodynamics. Having a greater range
of electron density data would also stop the ion
density data going to zero, at distances greater
than 5Rʘ; hence also stopping the magnetic field
strength and magnetic tension force going to
zero.
It must also be noted that between 2Rʘ and
5Rʘ, a full wavelenth was only witnessed on a
small number of the images during the event,
such as figure 15. In figures 15 and 16, the red line indicates the streamer that was analysed, and the
red diamond indicates where the CME originated. Once the oscillation was set into motion and started
to travel along the streamer, the oscillation very quickly passed beyond 5Rʘ, as so could no longer be
Figure 15: An image showing the oscillation along the streamer on 21/04/2013 from 07:12 to 11:00.
Figure 16: An image showing the oscillation along the streamer on 23/06/2012 from 15:48 to 16:24.
Nicholas David Mellor Streamer Waves 130053235
27
measured. For the earlier images, where there was no complete wavelengh, it has to be assumed that
the half wavelength that can be seen, would be symmetrical to the half that has not yet formed. This
assumption is essential to the calculations, as without it, the wavenumber, the frequency of the wave,
and hence the time period cannot be calculated reliably. Another factor that was found to determine
the wavelength of the streamer oscillation, was how far radially the streamer was from the source of
the CME. For example, the streamer shown in figure 16 was further away from where the CME
originated, than the streamer in figure 15 was. It was seen that the streamer in figure 16 started to
oscillate noticably at 3Rʘ, and the oscillations would extend out beyond 5Rʘ. However, it was also
noted that the wavelength of this oscillation, was shorter than the wavelength of the streamer
oscillation in figure 15, that was closer to the origin of the CME.
Comparing the results from this experiment with the results presented in the Astrophysical
Journal (Chen, Song et al., 2010); it can be seen in figure 17, that the wave speeds calculated in this
experiment are substantially higher than the literature values, and increases with distance rather than
decreases. This discrepancy is most likely due to the way the wavelength is calculated in this
experiment, the problems of which have already been discussed. However, since this paper was
published, the temporal resolution of LASCO-C2 observations have decreased from approximately 30
minutes, to 10. 6̇ and 12.8 minutes. This decrease in the time interval between observations could also
cause the discrepancy in the two sets of results.
Calculated wave speed (kms-1) 643 888 801 813 1273
Literature wave speed (kms-1) (Chen, Song et al., 2010)
429 391 344 369 325
Figure 17: A table comparing calculated and literature wave speeds.
After the initial results were gathered and it was seen that the wave speeds were much higher than
the literature values, further results were gathered to be compared. Figure 18 shows these results.
Attempt 1 in the table below is the set of results presented in this paper, and then 4 further attempts
were made. It can be seen from this table that the results from this experiment are consistently higher
than the literature values. Despite a few anomalous values such as the last value of the first attempt,
and the middle three values of attempt two.
Average mid-distance 53.27 66.09 77.49 91.30 103.08
Attempt 1 Wave speed (km/s) 643 888 801 813 1273
Attempt 2 Wave speed (km/s) 559 679 594 607 854
Attempt 3 Wave speed (km/s) 522 769 718 860 919
Attempt 4 Wave speed (km/s) 597 867 706 720 894
Attempt 5 Wave speed (km/s) 561 673 719 716 799 Figure 18: A table to show further wave speed results.
5.3. Error Analysis After considering figure 18, it can be seen that the wave speeds cannot be the main source of
error throughout this experiment. While some errors do occur from these values, it is not enough to
cause the inconsistencies seen in the magnetic field calculations. As such, other sources had to be
considered. When looking at the equation for the magnetic field strength, 𝐵0 = 𝑣𝐴√(𝜇0𝜌), it can be
seen that the only variables in the equation are the Alfvén speed, and the plasma density. It has already
been stated that the Alfvén speed is related to the wave speed by, 𝑣𝐴 =𝑐𝐾
√2, so the Alfvén speed will
also not be the major cause of the errors. This leaves the plasma density, when the plasma density is
plotted for the angle required against the heliocentric height as in figure 19, it can be seen that there
is no definite correlation.
Nicholas David Mellor Streamer Waves 130053235
28
Aside from the lack of a consistent correlation with heliocentric height, it can be seen that there are
also negative values of electron density. However, this is not logical, as there cannot be negative
density values, and as the heliocentric height increases the density value should decrease. It is these
factors that cause the discrepancy in the magnetic field. Without this consistent correlation, the
magnetic field strength values calculated from these values will not have a reliable correlation.
However, the magnetic field strength values are correct for the electron density values that have been
returned. It must also be noted that from figure 18, the mid-distance points start from approximately
50 on the x-axis, so any data before an x-axis value of 50 can be dismissed.
Discrepancies in the electron density
data could arise from the model that is used
to generate the data (Quémerais and Lamy,
2002). Whilst this model was being
implemented it could have been possible
that there were large variations within the
coronal plasma density itself. These
variations would then be shown in the
electron density data. Therefore, it will be
variations such as these that cause the errors
within the calculations.
After these initial results found that
the electron density model is the main
source of error throughout this experiment,
a new method for gathering the electron
density was devised, this is shown in
appendix 3. The radial line that has been
used so far is 578, along this line at each mid-
distance point the electron density value was
taken. In the revised method, two additional radial lines where taken either side of the main radial
Figure 19: A graph showing the electron density along the normalised heliocentric height, at an angle of 578o of the data structure.
Figure 20: A graph to show the average ion plasma density with heliocentric height.
Nicholas David Mellor Streamer Waves 130053235
29
line; e.g. 576, 577, 579 and 580. Along each of these lines, the electron density values were taken for
each mid-distance point; these were then averaged and manipulated to find an average ion density.
The difference can be seen in figure 20 when compared to figure 11. It can be seen from figure 21,
which compares the numerical data for figures 11 and 20, that the values of the averaged ion density
are mostly lower than the ion plasma density, except the values at 4.47Rʘ. Hence, it is thought that
using an averaged ion density set would increase the accuracy of the results.
Heliocentric height (R_sun)
Ion plasma density (m-3)
Average ion plasma density (m-3)
3.09 1.86E-15 1.61E-15
3.49 1.49E-15 1.47E-15
3.94 1.32E-15 1.29E-15
4.47 7.50E-16 7.75E-16
5.17 0 0 Figure 21: A table comparing the ion densities to the average ion densities.
Nicholas David Mellor Streamer Waves 130053235
30
6. Conclusion and Acknowledgements
6.1. Conclusion The main aim of this experiment was to calculate the properties, of an oscillating streamer
after a collision with the shock front from a coronal mass ejection. This was achieved successfully with
IDL and Solarsoft, in the form of a program. Two streamer/ shockwave collision events were examined,
on the 23/06/2012 from 15:48 to 16:24 and the 21/04/2013 from 07:12 to 11:00. However, due to
limitations with electron plasma density, the event on 23/06/2012 from 15:48 to 16:24, had to be
dismissed. All the calculations were therefore only carried out on the event on 21/04/2013 from 07:12
to 11:00. After analysing the results that were gathered from this event, it can be seen that the wave
speed and Alfvén speed increases with heliocentric height; ion plasma density, magnetic field strength
and the magnetic tension force, all decrease with heliocentric height. It was also found that the
frequency and time period of the wave decreases with heliocentric height, this would suggest a
damping effect. However, it is the magnetic tension force that acts as the damping force. The magnetic
tension force is the force that straightens bend magnetic field lines.
This method proves to be fairly successful, despite the subjective nature of it, it relies on the
user’s own perception of where boundaries are within the images. The IDL program written, uses the
Solarsoftware (ssw) environment, which provides a library of procedures written specifically for LASCO,
among other instruments. It is believed that this method is a good starting point to be built upon to
produce a more robust, and accurate method for calculating these properties.
The results of this experiment were compared to an earlier study in the Astrophysical Journal
(Chen, Song et al., 2010). After the comparison had been made it was found that the wave speed
results were vastly different and had different trends. Possible explanations for this were discussed in
the above section, however, as different methods and different streamer/ shockwave collision events
were used, the results would be different regardless.
6.2. Acknowledgements I would like to thank Dr. Xing Li from the Institute of Maths, Physics and Computer Science
(IMPACS) at Aberystwyth University, without his help and guidance this project would not have been
possible. I would also like to thank Dr. Huw Morgan also from the Institute of Maths, Physics and
Computer Science (IMPACS) at Aberystwyth University. For without the electron density data that he
supplied the project would have become much, much harder. Finally, I would like to thank the authors
who contributed to the IDL libraries used, such as Solarsoft, again without their contribution, the
project would not have come to fruition.
Nicholas David Mellor Streamer Waves 130053235
31
7. Additional Literature This section contains any literature that was used after the completion of the literature
search.
7.1. Bibliography
(NRL), N.R.L., Large Angle and Spectrometric Coronagraph Experiment. (NRL), N.R.L., LASCO/EIT Images Query Form. Aschwanden, M.J., 2005, Physics of the Solar Corona. An Introduction with Problems and Solutions
(2nd edition). Aschwanden, M.J., 2006, Philosophical Transactions of the Royal Society of London A: Mathematical,
Physical and Engineering Sciences 364(1839): 417. Burlaga, L.F., Plunkett, S.P., & St. Cyr, O.C., 2002, Journal of Geophysical Research: Space Physics
107(A10): SSH 1. Center, G.S.F., The IDL Astronomy User's Library. Chen, Y., 2013, Chin. Sci. Bull. 58(14): 1599. Chen, Y., Song, H.Q., Li, B., et al., 2010, The Astrophysical Journal 714(1): 644. Consulting, F.S., Coyote's guide to IDL programming. Corporation, H., IDL. Domingo, V., Fleck, B., & Poland, A., 1995, Space Science Reviews 72(1-2): 81. Doorsselaere, T.V., Nakariakov, V.M., & Verwichte, E., 2008, The Astrophysical Journal Letters 676(1):
L73. Freeland, S., & Handy, B., 1998, Solar Physics 182(2): 497. Freeland, S., L. Bentley, R, D, Solarsoft. G. E. Brueckner, e.a., 1995, Solar Physics 162(1). Golub, L., & Pasachoff, J.M., 2010, The solar corona. (Cambridge University Press). Gosling, J.T., Bame, S.J., McComas, D.J., et al., 1994, Geophysical Research Letters 21(12): 1109. Gosling, J.T., Birn, J., & Hesse, M., 1995, Geophysical Research Letters 22(8): 869. Guenther, D., 1989, The Astrophysical Journal 339: 1156. Hasegawa, A., & Uberoi, C., 1982, Alfven wave. DOE Critical Review Series, (Medium: ED; Size: Pages:
137). Howard, R.A., Sheeley, N.R., Koomen, M.J., & Michels, D.J., 1985, Journal of Geophysical Research:
Space Physics 90(A9): 8173. Hudson, H.S., Lemen, J.R., St. Cyr, O.C., Sterling, A.C., & Webb, D.F., 1998, Geophysical Research Letters
25(14): 2481. Hundhausen, A.J., 2012, Coronal expansion and solar wind. (Springer Science & Business Media). J. T. Gosling, P.R., 1996, Geophysical Research Letters 23(21). K. L. DeRose, C.C., P. Pribyl, W. Gekelman, 2008, Measurement of the Plasma Beta in the Enormous
Toroidal Plasma Device (ETPD), (UCLA, Department of Physics and Astronomy - University of Califonia, LA, CA, 90095-1547).
Koutchmy, S., & Livshits, M., 1992, Space Science Reviews 61: 393. Markus, J., 2004, Physics of the Solar Corona: An Introduction, (Springer-Verlag, Berlin). Mathioudakis, M., Jess, D.B., & Erdélyi, R., 2013, Space Science Reviews 175: 1. Morgan, H., 2015, The Astrophysical Journal Supplement Series 219(2): 23. Moussas, X., Polygiannakis, J., Preka-Papadema, P., & Exarhos, G., 2005, Advances in Space Research
35(5): 725. Nakariakov, V.M., Melnikov, V.F., & Reznikova, V.E., 2003, Astronomy and Astrophysics 412: L7. Nakariakov, V.M., & Verwichte, E., 2005, Living reviews in solar physics 2: 3. NASA, Index of /CME_list/daily_movies. Osterbrock, D.E., 1961, The Astrophysical Journal 134: 347.
Nicholas David Mellor Streamer Waves 130053235
32
Pence, W.D., Chiappetti, L., Page, C.G., Shaw, R., & Stobie, E., 2010, Astronomy & Astrophysics 524: A42.
Phillips, K.J., 1995, Guide to the Sun. (Cambridge University Press). Priest, E., 2014, Magnetohydrodynamics of the Sun. (Cambridge University Press). Priest, E., Forbes, T., & Forbes, T., 2000, Magnetic Reconnection. Quémerais, E., & Lamy, P., 2002, Astronomy & Astrophysics 393(1): 295. Raymond, J.C., Kohl, J.L., Noci, G., et al., 1997, Solar Physics 175(2): 645. Raymond, J.C., Thompson, B.J., St. Cyr, O.C., et al., 2000, Geophysical Research Letters 27(10): 1439. Ridpath, I., 2012, A dictionary of astronomy. (Oxford University Press). Roberts, B., & Webb, A.R., 1978, Solar Physics 56(1): 5. Santamaria, I.C., Khomenko, E., & Collados, M., 2015, Astronomy and Astrophysics 577: 70. Spruit, H.C., 1982, Solar Physics 75(1-2): 3. Steiner, O., 2006. Stern, B., 2000, Proceedings of SPACE: 1011. Subramanian, P., Dere, K.P., Rich, N.B., & Howard, R.A., 1999, Journal of Geophysical Research: Space
Physics 104(A10): 22321. U. Feldman, E.L., N. A. Schwadron, 2005, Journal of Geophysical Research: Space Physics 110. V. Bothmer, R.S., 1996, Advances in Space Research 17(4). V. Yurchyshyn, S.Y., H. Wang, N. Gopalswamy, Statistical Distributions of Speeds of Coronal Mass
Ejections, (Big Bear Observatory, Centre for Solar Physics ans Space Weather, Crimean Astrophysical Observatory, Centre for Solar Research of New Jersey Institute of Technology, Laboratory for Experimental Physics, Califonia, USA. Washington, USA. Crimea, Ukraine. New Jersey, USA. Maryland, USA).
Wang, Y.-M., Biersteker, J., Sheeley Jr, N., et al., 2007, The Astrophysical Journal 660(1): 882. Wells, D.C., Greisen, E., & Harten, R., 1981, Astronomy and Astrophysics Supplement Series 44: 363. Withbroe, G.L., & Noyes, R.W., 1977, Annual Review of Astronomy and Astrophysics 15(1): 363. Zhao, X., & Webb, D., 2003, Journal of Geophysical Research: Space Physics 108(A6). Zirker, J.B., 2002, Journey from the Center of the Sun. (Princeton University Press).
Nicholas David Mellor Streamer Waves 130053235
33
8. Appendices
8.1. Appendix 1 The data within this appendix is the tabulated form of the graphs given in the results section.
The graph provided here in figure 24, shows relationship between wave frequency and heliocentric
height, but this is not directly needed for the experiment.
Height from solar
centre (R_sun)
Streamer wavelength
(m)=
Wave speed
(km/s)=
Alfven speed
(km/s)=
Ion plasma density
(m-3)
Magnetic field
strength (T)
Magnetic tension
force (Pa/m)
Angular frequency (rads/s)=
Time period
(hours)=
3.08839 1.79E+06 643 455 1.86E-15 2.20E-05 1.93E-04 2.26E-03 7.72E-01
3.49369 2.28E+06 888 628 1.49E-15 2.72E-05 2.94E-04 2.45E-03 7.12E-01
3.93747 2.72E+06 801 566 1.32E-15 2.31E-05 2.12E-04 1.85E-03 9.43E-01
4.46815 3.39E+06 813 575 7.50E-16 1.76E-05 1.24E-04 1.50E-03 1.16E+00
5.17242 3.83E+06 127 900 0 0 0 -NaN -NaN Figure 22: A table displaying all the data gathered during the experiment.
Visible streamer length (km) 2.71E+06
Average streamer Wavelength (m) 2.80E+06 Figure 23: A table showing some average results.
Figure 24: A graph showing the wave frequency against heliocentric height.
Nicholas David Mellor Streamer Waves 130053235
34
8.2. Appendix 2
This appendix displays the code used to calculate the required results; the code varies
depending on the data set used. However, this appendix shows the code required for the 21/04/2013
dataset.
PRO mainprogram_21042013
DEVICE, Decomposed=0
LOADCT,0
!p.multi = 0
!p.background = 0
;;;changing the directory to find the images
dir='C:\Users\Nick\Documents\Uni_stuff\Third_year\Disseration\Fits_img\2104
201305'
;;;search folder for number of files
filesearch=FILE_SEARCH(dir, '*', COUNT=n)
PRINT,'number of files=', n
;WINDOW, n, xsize=1250, ysize=725
;;;finding the files
PRINT,findfile(dir+'\*.fts')
file=findfile(dir+'\*.fts')
;;;Giving the files a name
imgs=INTARR(1024,1024,n)
obvtime=STRARR(n)
exptime=FLTARR(n)
arcsec=FLTARR(n)
datee=STRARR(n)
FOR i=0,n-1 DO BEGIN
imgs[*,*,i]=LASCO_READFITS(file(i),headerinfo)
obvtime(i)= headerinfo.TIME_OBS
exptime(i)= headerinfo.EXPTIME
arcsec(i)=headerinfo.PLATESCL
datee(i)=headerinfo.DATE_OBS
ENDFOR
;;;to calculate time
tsec=fltarr(n)
tsecbetween=fltarr(n)
FOR i=0,n-1 DO BEGIN
tsec(i)=ANYTIM2TAI(ANYTIM2CAL(datee(i)+' '+obvtime(i),form=11))
ENDFOR
FOR i=0,n-2 DO BEGIN
tsecbetween(i)=tsec(i+1) - tsec(i)
ENDFOR
PRINT,arcsec
PRINT,datee
PRINT,obvtime
;;;setting up the image plots, giving the orientation in the window.
;;;incase of odd number of n
WINDOW,n, xsize=1024, ysize=725
m=EVENODD(n) ;m=0 when n is even, m=1 if n is odd
IF m EQ 1 THEN BEGIN
o=((n/2.0)+0.5)
ENDIF ELSE BEGIN
o=n/2
ENDELSE
!P.multi=[0,o,2,1024] ; 3X2 multi-plots
!P.CHARSIZE=1.
!X.CHARSIZE=1.
Nicholas David Mellor Streamer Waves 130053235
35
!Y.CHARSIZE=1.
;;;plotting the images.
FOR i=0,n-1 DO BEGIN
PLOT_IMAGE,imgs[*,*,i], title='Image'+STRING(i)
ENDFOR
;;;plotting each of the images in their own window and making them all
512x512
images=INTARR(1024,1024,n)
!P.multi=0
minim=MIN(imgs,dim=3)
x=FINDGEN(1024)-headerinfo.crpix1
y=FINDGEN(1024)-headerinfo.crpix2
xx=REBIN(x,1024,1024)
yy=REBIN(REFORM(y,1,1024),1024,1024)
ht=SQRT(xx^2+yy^2)
mask=ht GT 175
HELP, images
FOR i=0,n-1 DO BEGIN
WINDOW,i,xsize=1024,ysize=1024
images[*,*,i]=HIST_EQUAL((imgs[*,*,i]-minim)*mask,per=0.01)
TVSCL,images[*,*,i];,/nosquare, xma,rgin=[0,0], ymargin=[0,0]
ENDFOR
sizee=SIZE(images)
rdi=INTARR(sizee(1),sizee(2),n)
rdisize=(SIZE(rdi[*,*,1]))
rdiresize=INTARR(rdisize(1)*3,rdisize(2)*3,n)
for i=1,n-2 DO BEGIN
rdi[*,*,i]=images[*,*,(i+1)]-images[*,*,i]
; tvscl,rdi[*,*,i];,/nosquare, ymargin=[0,0]
rdiresize[*,*,i]=CONGRID(rdi[*,*,i],rdisize(1)*3,rdisize(2)*3,/interp)
endfor
!P.multi=0
PRINT, 'click at each node'
;;;calculating the approximate length of streamer
WINDOW,n+1,xsize=1024,ysize=1024
!P.multi=[0,1,1,612]
TVSCL,images[*,*,3]
PRINT, 'Select the lower right hand point and the upper left hand point
of the wavelength, to be calculated.'
PRINT, 'Change in the code if different dimensions are necessary.'
CURSOR, X1,Y1, /DEVICE
WAIT, 1s
CURSOR, X2,Y2, /DEVICE
x_b=ABS(X1-X2)
y_b=ABS(Y2-Y1)
xy_b=SQRT((x_b)^2+(y_b)^2) ;streamer in
pixels
wlarcsec=xy_b*arcsec[2] ;streamer in
arcsecs
stlength=wlarcsec*730 ;streamer
length in km
stlength=stlength+(696000)
;streamer length + 1 solar radii
PRINT, 'streamer length in arcsecs', wlarcsec
PRINT, 'Streamer length (km)=', stlength
Nicholas David Mellor Streamer Waves 130053235
36
;;;to calculate the wavespeed of the streamer
a=n-4
b=4
wavespeed=FLTARR(5)
alfvenspeed=FLTARR(5)
middistance=FLTARR(5)
middistance1=FLTARR(5)
WINDOW,n+2, xsize=1024, ysize=1024
FOR i=4, n-6 DO BEGIN
TVSCL,rdi[*,*,i]
PRINT,'Select two nodes of waves'
WAIT,1s
CURSOR, X4,Y4,/DEVICE
WAIT, 1s
CURSOR, X5,Y5,/DEVICE
X3=headerinfo.crpix1
Y3=headerinfo.crpix2
XX=ABS(X4-X3)
YY=ABS(Y4-Y3)
XY=SQRT((XX)^2+(YY)^2) ;distance
from centre of sun to first wave node
XXX=ABS(X5-X3)
YYY=ABS(Y5-Y3)
XY2=SQRT((XXX)^2+(YYY)^2) ;distance
from centre of sun to second wave node
xydiff=ABS(XY2-XY)
;difference between the first and second wave node
xydiffarcsec=xydiff*arcsec[2]
;difference in arcsecs
difference=xydiffarcsec*730
;sdifference in km
middistance(i-4)=(((((xydiff/2)+xy)*arcsec[2]*730)-696342.0)/26112.825)
;distance from limb to midpoint
middistance1(i-4)=((((xydiff/2)+xy)*arcsec[2]*730)/696342)
wavespeed(i-4)=(difference/(tsecbetween(i)))*1000
alfvenspeed(i-4)=wavespeed(i-4)/(SQRT(2.0))
IF wavespeed(4) GT 0 THEN BREAK
ENDFOR
PRINT, 'wavespeed',wavespeed
PRINT, 'Alfven speed', alfvenspeed
PRINT, 'middistance', middistance
stop
averagewavespeed=(TOTAL(wavespeed))/(N_ELEMENTS(wavespeed))
wavelength=FLTARR(5)
frequency=FLTARR(5)
timeperiod=FLTARR(5)
FOR i=4,n-6 DO BEGIN
;;;Calculate the wavelengths
TVSCL,rdi[*,*,i]
Nicholas David Mellor Streamer Waves 130053235
37
PRINT,'Select two crests of waves'
WAIT,1s
CURSOR, X6,Y6,/DEVICE
WAIT, 1s
CURSOR, X7,Y7,/DEVICE
XX=ABS(X6-X3)
YY=ABS(Y6-Y3)
XY=SQRT((XX)^2+(YY)^2) ;distance
from centre of sun to first wave node
XXX=ABS(X7-X3)
YYY=ABS(Y7-Y3)
XY2=SQRT((XXX)^2+(YYY)^2) ;distance
from centre of sun to second wave node
xydiff=ABS(XY2-XY)
;difference between the first and second wave node
xydiffarcsec=xydiff*arcsec[2]
;difference in arcsecs
difference=xydiffarcsec*730
;sdifference in km
wavelength(i-b)=(difference)*2
frequency(i-b)=wavespeed(i-b)/wavelength(i-b)
timeperiod(i-b)=(1.0/frequency(i-b))
IF wavelength(4) GT 0 THEN BREAK
ENDFOR
k=FLTARR(5)
;wavenumber
pi=3.141592654
FOR i=0,4 DO BEGIN
k(i)=(2*pi)/(wavelength(i)*1000)
PRINT,'Angular wavenumber=', k(i)
ENDFOR
;;;radius of curvature
R_c=FLTARR(5)
h=FLTARR(5)
w=FLTARR(5)
FOR i=4,n-6 DO BEGIN
WINDOW, n+3, xsize=1024,ysize=1024
TVSCL, images[*,*,i]
PRINT,'First click'
WAIT,1
CURSOR, X8,Y8,/device,/wait
WAIT, 1
PRINT,'second click'
CURSOR, X9,Y9,/device,/wait
PLOTS, [X8,X9], [Y8,Y9], /DEVICE
WAIT,1
PRINT,'third click'
CURSOR, X10,Y10,/DEVICE,/wait
;stop
WAIT, 1
PRINT,'fourth click'
CURSOR, X11,Y11,/DEVICE,/wait
WAIT,1
XX=ABS(X8-X3)
YY=ABS(Y8-Y3)
Nicholas David Mellor Streamer Waves 130053235
38
XY=SQRT((XX)^2+(YY)^2) ;distance
from centre of sun
XXX=ABS(X9-X3)
YYY=ABS(Y9-Y3)
XY2=SQRT((XXX)^2+(YYY)^2) ;distance
from centre of sun to second
xydiff=ABS(XY2-XY) ;difference
between the first and second wave node
xydiffarcsec=xydiff*arcsec[2] ;difference
in arcsecs
w(i-4)=xydiffarcsec*730 ;sdifference in
km
;;;;;;;;;;;;;;;;;;;;;;;
XX=ABS(X10-X3)
YY=ABS(Y10-Y3)
XY=SQRT((XX)^2+(YY)^2) ;distance
from centre of sun
XXX=ABS(X11-X3)
YYY=ABS(Y11-Y3)
XY2=SQRT((XXX)^2+(YYY)^2) ;distance
from centre of sun to second
xydiff=ABS(XY2-XY) ;difference
between the first and second wave node
xydiffarcsec=xydiff*arcsec[2] ;difference
in arcsecs
h(i-4)=xydiffarcsec*730 ;sdifference in
km
R_c(i-4)=(h(i-4)/2.0)+(w(i-4)^2)/(8.0*h(i-4))
ENDFOR
;;;Electron Density
dir2='C:\Users\Nick\Documents\Uni_stuff\Third_year\Disseration\ElectronDens
ity'
a=DIALOG_PICKFILE(path=dir2);choose file
RESTORE,a,/verbose ; read file
HELP,d ;file is restored into variable “d”, which is a data structure.
;TVSCL,d.nel ; display electron density
;TVSCL,d.nelf ; display a more smoothed (less noise) version of electron
density
;PLOT,d.nelf[*,20],/xst
;plot a latitudinal slice at a given height
;OPLOT,d.nel[*,20],thick=2
;print,'d.rra', d.rra ;print the height range of the polar-coordinate
array
hts=d.rra[0]+FINDGEN(d.nht)*(d.rra[1]-d.rra[0])/(d.nht-1) ;create vector
of height coordinates
pa=d.para[0]+FINDGEN(d.npa)*(d.para[1]-d.para[0])/d.npa ;create vector
of position angle coordinates (0 to 359)
rho=FLTARR(5)
rhoo=FLTARR(5)
middis=FLTARR(5)
FOR i=0,4 DO BEGIN
IF middistance(i) GT 99.0 THEN BEGIN
middistance(i)=0.0
ENDIF ELSE BEGIN
middis(i)=ROUND(middistance(i))
Nicholas David Mellor Streamer Waves 130053235
39
rho(i)=d.nel[578,middis(i)]
rhoo(i)=rho(i)*1.67E-27*1000000
ENDELSE
PRINT,'Plasma density=', rhoo(i)
ENDFOR
magneticfield=FLTARR(N_ELEMENTS(wavespeed))
mu0=1.2566E-6
FOR i=0, 4 DO BEGIN
magneticfield(i)= alfvenspeed(i)* SQRT(mu0*rhoo(i))
ENDFOR
averagemagnetic=(TOTAL(magneticfield))/(N_ELEMENTS(magneticfield))
averagewavelength=(TOTAL(wavelength))/(N_ELEMENTS(wavelength))
tensionforce=FLTARR(5)
FOR i=0,4 DO BEGIN
tensionforce(i)=(magneticfield(i)^2)/(2*mu0)+((magneticfield(i)^2)/(mu0))*(
1/(R_c(i)/1000))
ENDFOR
angfreq=FLTARR(5)
caltperiod=FLTARR(5)
FOR i=0,4 DO BEGIN
angfreq(i)=SQRT((2*(k(i)^2)*(magneticfield(i)^2))/(mu0*(rhoo(i))))
PRINT, 'angular frequency=',angfreq(i)
caltperiod(i)=((2*pi)/angfreq(i))/3600
PRINT, 'calculated time period (hours)=',caltperiod(i)
ENDFOR
!p.multi = 0
!p.background = 255
WINDOW,n+4, xsize=500, ysize=500
PLOT,middistance1,wavespeed,$
TITLE='Wave speed against height from solar centre',$
XTITLE='Height from solar centre (R_sun)', YTITLE='Wave speed (m/s)',$
THICK=1.0, /YNOZERO, PSYM=-7, COLOR=0, xmargin=[13,3]
WRITE_JPEG, FILEPATH('speedagainstheight.png',
ROOT_DIR='C:\Users\Nick\Documents\Uni_stuff\Third_year\Disseration\Data'),$
TVRD()
WINDOW,n+5, xsize=500, ysize=500
PLOT,middistance1,alfvenspeed,$
TITLE='Alfven speed against height from solar centre',$
XTITLE='Height from solar centre (R_sun)', YTITLE='Alfven speed
(m/s)',$
THICK=1.0, /YNOZERO, PSYM=-7, COLOR=0, xmargin=[13,3]
WRITE_JPEG, FILEPATH('alfvenagainstheight.png',
ROOT_DIR='C:\Users\Nick\Documents\Uni_stuff\Third_year\Disseration\Data'),$
TVRD()
;stop
WINDOW,n+6, xsize=500, ysize=500
PLOT,middistance1,magneticfield,$
TITLE='Magnetic field against height from solar centre',$
XTITLE='Height from solar centre (R_sun)', YTITLE='Magnetic field
strength (T)',$
THICK=1.0, /YNOZERO, PSYM=-7, COLOR=0, xmargin=[13,3]
WRITE_JPEG, FILEPATH('magfieldagainstheight.png',
ROOT_DIR='C:\Users\Nick\Documents\Uni_stuff\Third_year\Disseration\Data'),$
TVRD()
Nicholas David Mellor Streamer Waves 130053235
40
WINDOW,n+7, xsize=500, ysize=500
PLOT,middistance1,rhoo,$
TITLE='Ion plasma density against height from solar centre',$
XTITLE='Height from solar centre (R_sun)', YTITLE='Ion plasma density
(m^-3)',$
THICK=1.0, /YNOZERO, PSYM=-7, COLOR=0, xmargin=[13,3]
WRITE_PNG, FILEPATH('rhoagainstheight.png',
ROOT_DIR='C:\Users\Nick\Documents\Uni_stuff\Third_year\Disseration\Data'),$
TVRD()
WINDOW,n+8, xsize=500, ysize=500
PLOT,middistance1,tensionforce,$
TITLE='Magnetic tension force against height from solar centre',$
XTITLE='Height from solar centre (R_sun)', YTITLE='Magnetic tension
force (Pa/m)',$
THICK=1.0, /YNOZERO, PSYM=-7, COLOR=0, xmargin=[13,3]
WRITE_PNG, FILEPATH('tensionforceagainstheight.png',
ROOT_DIR='C:\Users\Nick\Documents\Uni_stuff\Third_year\Disseration\Data'),$
TVRD()
WINDOW,n+9, xsize=500, ysize=500
PLOT,middistance1,angfreq,$
TITLE='Wave frequency against height from solar centre',$
XTITLE='Height from solar centre (R_sun)', YTITLE='Frequency of the
wave (Rads/s)',$
THICK=1.0, /YNOZERO, PSYM=-7, COLOR=0, xmargin=[13,3]
WRITE_PNG, FILEPATH('freqagainstheight.png',
ROOT_DIR='C:\Users\Nick\Documents\Uni_stuff\Third_year\Disseration\Data'),$
TVRD()
PRINT, 'Height from solar centre (R_sun)=',middistance1
PRINT, 'Time between images (s)', tsecbetween
PRINT, 'Average wavespeed (m/s) =',wavespeed
PRINT, 'Alfven speed (m/s)=',alfvenspeed
PRINT, 'Wavelength (km)=',wavelength
PRINT, 'Magnetic field strength (T)=', magneticfield
PRINT, 'Max Amplitude=',h
PRINT, 'w=',w
PRINT, 'Radius of curvature (km)=',R_c
PRINT, 'Magnetic Tension force (T)=', tensionforce
PRINT, 'Angular frequency (rads/s)=', angfreq
PRINT, 'Time period (hours)=', caltperiod
dir2='C:\Users\Nick\Documents\Uni_stuff\Third_year\Disseration\Data\'
OPENU,1, dir2+'STREAMER_DATA.dat'
PRINTF,1, 'Streamer length (km)=', stlength
PRINTF,1, 'Streamer wavelength (m)=', wavelength
PRINTF,1, 'Wave speed (m/s)=',wavespeed
PRINTF,1, 'Average streamer Wavelength (m)', averagewavelength
PRINTF,1, 'Height from solar centre (R_sun)=',middistance1
PRINTF,1, 'Time between images (s)', tsecbetween
PRINTF,1, 'Alfven speed (m/s)=', alfvenspeed
PRINTF,1, 'Ion plasma density=', rhoo
PRINTF,1, 'Magnetic field strength (T)',magneticfield
PRINTF,1, 'Magnetic Tension force (Pa/m)=', tensionforce
PRINTF,1, 'Angular frequency (rads/s)=', angfreq
PRINTF,1, 'Time period (hours)=', caltperiod
CLOSE,1
STOP
END
Nicholas David Mellor Streamer Waves 130053235
41
8.3. Appendix 3 This appendix contains the revised bit of code to calculate the average ion density. This
replaces the initial code finding the original ion density.
rho=FLTARR(5)
rho1=FLTARR(5)
rho2=FLTARR(5)
rho3=FLTARR(5)
rho4=FLTARR(5)
rhoo=FLTARR(5)
rhoo1=FLTARR(5)
rhoo2=FLTARR(5)
rhoo3=FLTARR(5)
rhoo4=FLTARR(5)
middis=FLTARR(5)
FOR i=0,4 DO BEGIN
IF middistance(i) GT 99.0 THEN BEGIN
middistance(i)=0.0
ENDIF
middis(i)=ROUND(middistance(i))
rho(i)=abs(d.nel[576,middis(i)])
rho1(i)=abs(d.nel[577,middis(i)])
rho2(i)=abs(d.nel[578,middis(i)])
rho3(i)=abs(d.nel[579,middis(i)])
rho4(i)=abs(d.nel[580,middis(i)])
rhoo(i)=rho(i)*1.67E-27*1000000
rhoo1(i)=rho1(i)*1.67E-27*1000000
rhoo2(i)=rho2(i)*1.67E-27*1000000
rhoo3(i)=rho3(i)*1.67E-27*1000000
rhoo4(i)=rho4(i)*1.67E-27*1000000
;PRINT,'Plasma density=', rhoo(i)
ENDFOR
averagerho=FLTARR(5)
averagerhoo=FLTARR(5)
FOR i=0,4 DO BEGIN
averagerho(i)=(rho(i)+rho1(i)+rho2(i)+rho3(i)+rho4(i))/5
averagerhoo(i)=(rhoo(i)+rhoo1(i)+rhoo2(i)+rhoo3(i)+rhoo4(i))/5
ENDFOR