The delta-spot growth and formation of NOAA active region 7978 and its relation to nonlinear numerical simulations models

Material Information

The delta-spot growth and formation of NOAA active region 7978 and its relation to nonlinear numerical simulations models
Meisner, Randle W
Publication Date:
Physical Description:
xi, 80 leaves : illustrations ; 29 cm


Subjects / Keywords:
Sunspots ( lcsh )
Sunspots ( fast )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references (leaves 76-80).
General Note:
Submitted in partial fulfillment of the requirements for the degree, Master of Basic Science.
Statement of Responsibility:
by Randle W. Neisner.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
37854014 ( OCLC )
LD1190.L44 1997m .M45 ( lcc )

Full Text
Randle W. Meisner
B. A., University of Colorado at Boulder, 1991
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Basic Science

This thesis for the Master of Basic Science
degree by
Randle W. Meisner
has been approved
Martin E. Huber
/ ^ Date

Meisner, Randle W. (Master of Basic Science)
The Delta-Spot Growth and Formation of NOAA Active Region 7978 and
Its Relation to Nonlinear Numerical Simulation Models
Thesis directed by Professor Clyde Zaidins
The growth and formation of the delta-sunspot group NOAA AR 7978 is
analyzed with white light, hydrogen alpha, and soft X-ray images in this
study. Sunspot proper motions are determined from the white light
images for the dominant spots of this active region, as well as for the
umbrae involved with the deltaconfiguration, and show no extraordinary
rate variations from studies performed in the past. It is concluded that
the motions of one of the deltaspot umbrae and a small exterior pore
exhibit a direct magnetic link and indicate the possible ascent of a closed
magnetic loop. The white light images are also used to determine the
change in umbral area with the formation of penumbrae around two of
the sunspots. However, since only two sunspots are analyzed, no
conclusions are drawn about the relationship between umbral area and
the formation of a penumbra. The hydrogen alpha and soft X-ray images
are analyzed to test the predictions of recently published nonlinear
numerical simulations on the ascent of an active region magnetic flux
tube through the photosphere. It is shown here that soft X-ray bright
points and the disappearance of a hydrogen alpha filament are both
signatures of new emerging magnetic flux, but they are not necessarily
simultaneous in space and in time as the numerical simulations predict.
This abstract accurately represents the content of the candidates thesis.
I recommend its publication.
Clyde Zaidins

I dedicate this thesis to my very loving and patient wife, Brooke, who has
endured my long hours of study, and sometimes frustration, for the past
three years.

My many thanks to the members of my graduate committee for their
advice and support, to Dr. Kiplinger for his assistance in acquiring and
processing SOONSPOT data, to Dr. Shibata for the use of Yohkoh soft X
ray data, and to Dr. Karen Harvey for providing me with Kitt Peak

Tables........................................... xi
1. INTRODUCTION.........................................1
Purpose of This Study..........................2
Scope of the Data..............................3
Arrangement of the Thesis......................4
Solar Cycle....................................9
Proper Motion of Sunspots................10
Delta Sunspots...........................11
The Physics of Modeling Sunspots.........13
Magnetic Structure of Sunspots...........17
Solar Flares..................................20

3. INSTRUMENT DESCRIPTIONS............................24
Kitt Peak....................................27
4. WHITE LIGHT OBSERVATIONS...........................29
Sunspot Proper Motions.......................30
Analysis Technique......................32
Computing Solar Coordinates........32
Determining Sunspot Centers........35
Proper Motions of the Major Sunspots....36
Proper Motion of the Delta-Spot.........40
Penumbrae Formation..........................44
Measurements of Penumbral Formation.....45
Results of Penumbral Formation..........48
SOFT X-RAY OBSERVATIONS............................51
Numerical Simulations and Models.............52
Numerical Methods and Predictions.......52
Testing of Nonlinear Numerical Simulations...55
Selection of Events.....................56

Co-alignment of Images...................57
Test Results.............................59
6. DISCUSSION AND CONCLUSION............................69
White Light Observations.......................70
Sunspot Proper Motions...................70
Delta-Spot Proper Motions................71
Penumbral Formation......................72
Hydrogen Alpha and Soft X-Ray Observations.....74
Closing Statement..............................75

2.1 Properties of the Sun...................................6
2.2 Sunspot General Features................................8
2.3 Solar Cycle............................................10
2.4 Magnetically Split Spectral Line.......................14
4.1 AR 7978 Sunspot Labeling...............................31
4.2 Solar Disk (x, y) Coordinate System....................33
4.3 Example of Umbral Contours.............................36
4.4 Dominant Sunspot Proper Motions........................38
4.5 Kitt Peak Magnetograms.................................39
4.6 Delta Spot Proper Motions..............................41
4.7 White Light DeltaSpot Time Series.....................42
4.8 Delta Spot Flux Tube Sketch............................43
4.9 Intensity Histograms of Sunspot B......................46
4.10 Intensity Histograms of Sunspot D......................47
4.11 Penumbral Formation Sketches...........................49
5.1 Sketch of Numerical Simulation Models..................54
5.2 White Light Images for Numerical Simulation Test.......60
5.3 Hydrogen Alpha Image Series of Event One...............61
5.4 Soft X-Ray Images of Event One.........................62
5.5 Hydrogen Alpha Image Series of Event Two...............63

5.6 Soft X-Ray Images of Event Two.........................64
5.7 Hydrogen Alpha Image Series of Event Three.............66
5.8 Soft X-Ray Images of Event Three.......................67
6.1 Evolution of X Class Flare.............................73

2.1 Polarization Properties of Spectra........................15
3.1 Definitions of SOONSPOT Header Items......................25
4.1 Umbral Area Measurements..................................48

Sunspots and solar active regions have been studied since the time
of Galileo in the early 1600s. However, it was not until the beginning of
the twentieth century when George Ellery Hale discovered that sunspots
were localized regions of intense magnetic fields. Since then, scientists
have been attempting to model and understand the Suns magnetic field
and its effects on the solar atmosphere, interplanetary space, and
planetary environments.
It is important to gain an understanding of the physics of sunspots
and the Suns magnetic field because of the direct effects solar magnetic
disturbances can have on the Earth. The Suns magnetic field is directly
connected to the Earths magnetosphere, and the other planetary
magnetospheres, on the day side of the planet through open field lines. If
events in the solar atmosphere, such as flares, coronal mass ejections,
and proton events, occur along one of these connected field lines, the
disturbance can propagate along the field lines and impact the Earth.
Possible effects of such an impact could include loss of communications
and other satellite operations, health and safety hazards to astronauts
that may be in orbit, enhanced auroral activity, and perhaps even large
power outages in major cities. Therefore, it is important to study these
phenomena, not only for the sake of knowledge, but also for sustaining

our daily operations here on Earth.
Purpose of This Study
The purpose of this thesis is to analyze the rapid formation of a
complex sunspot group from July 1996. The active region (AR) is
referenced by the National Oceanic and Atmospheric Administration
(NOAA) as AR, 7978 and fully emerged over a period of a day and a half.
The focus of the study is on the initial emergence of magnetic flux
through the solar surface and utilizes data obtained for the period 7-9
July 1996.
The main objective of this study is threefold. First, I use white
light images of AR 7978 to determine sunspot proper motions as the
region develops. In particular, I study the umbral motions of the delta (8)
spot configuration which quickly forms on 8 July and persists for only one
day. I also use the white light images of the region to determine when,
relative to umbral diameter, sunspot penumbrae are formed and compare
the results to current theories and observations. Finally, I use co
aligned hydrogen alpha (H-a) and soft X-ray images to determine the
validity of two-dimensional (2D) and three-dimensional (3D) nonlinear
numerical simulations of emerging flux regions, published by Shibata et
al. (1989), Shibata, Nozawa, and Matsumoto (1992) and Matsumoto et al.

Scope of the Data
The data used in this thesis consists of hundreds of digital images
of AR 7978 in white light, H-a, soft X-rays, and photospheric
longitudinal magnetograms. The white light and H-a images were
acquired from the SOONSPOT (Solar Observing Optical Network, Solar
Patrol On Tape) network of telescopes on United States Air Force bases
around the world. The SOONSPOT network includes observatories in
Palehua, Hawaii, Holloman, New Mexico, Ramey, Puerto Rico, San Vito,
Italy, and Learmonth, Australia and is intended to provide uninterrupted
24 hour coverage of the Sun, weather permitting. However, in July 1996,
only three of the observatories, Holloman, Ramey, and San Vito, were
fully operational which resulted in only 12-13 hours of observations per
The soft Xray images were obtained from the Japanese Soft X-ray
Telescope (SXT) on board the Yohkoh satellite. Since observations during
this time period were not part of an organized campaign, the majority of
these images are at one-quarter of the instruments full resolution,
resulting in only moderate spatial resolution. Nevertheless, the image
quality of the data is adequate for this study. During the day side pass of
the satellites orbit, the temporal resolution of the SXT varies from a few
seconds up to 30 minutes, depending on flare activity, image resolution,
and telemetry. The night side pass of Yohkohs orbit results in a data gap
of approximately 33 minutes per orbit.
Full disk longitudinal magnetograms were provided by the Kitt
Peak National Solar Observatory outside of Tucson, Arizona. The

magnetograms taken at Kitt Peak are produced once per day and are
high resolution images acquired for both the photosphere and the
chromosphere. For the time period examined in this analysis, a
continuous set of images is only available for 4-8 July 1996, due to bad
weather conditions on following days. Since the temporal resolution of
this data set is very low, and only two of the three days in this study are
covered, only the photospheric magnetograms are used for reference of
the sunspot polarities and as indicators of newly emerged magnetic flux.
Arrangement of the Thesis
In what follows, I present a brief overview of the solar atmosphere,
and current theories and observations of solar active regions in chapter
two. In chapter three, the general aspects of the instruments used to
acquire the data presented in this thesis are discussed. Chapter four
contains my analysis of the sunspot proper motions of AR 7978 from the
SOONSPOT white light images, along with my findings of when sunspot
penumbrae develop relative to umbral diameter. In chapter five, I
present a summary of the numerical simulations investigated in this
analysis and describe the predictions that are tested. I also present my
H-a and soft X-ray data used to test these predictions of the nonlinear
numerical simulations and discuss my findings on the matter. Finally, in
chapter six, I discuss the rapid formation of AR 7978 and present my

The solar atmosphere can be generalized into four distinctive
layers: (1) the photosphere which is the surface of the Sun that we see
in white light; (2) the chromosphere which is the layer overlying the
photosphere and ranges in temperature from about 6000 K near the solar
surface to over 10,000 K in its outer regions; (3) the transition region
which is a thin shell surrounding the chromosphere where the
temperature increases abruptly to approximately lxlO6 K; and (4) the
corona which is the hottest and outer most structure of the Sun rising to
temperatures over 2xl06 K (Snow, 1987). Figure 2.1 is a graphical
representation of solar atmosphere and includes values of some of the
basic properties of each region.
The main characteristic of the Sun that is prominent throughout
its atmosphere is the solar magnetic field. The overall average magnetic
field of the Sun is on the order of a few gauss, or a few times 1CH tesla
(Zirin, 1988; Parks, 1991), compared to the Earths value of about 0.3
gauss at the equator (Parks, 1991). However, the solar magnetic field is
not constant, nor is it homogeneous. On scales down to thousands, or
maybe even hundreds, of meters the magnetic structure of the Sun varies
considerably from 1 to 3,000 gauss in extreme cases. The location of
these areas with strong magnetic gradients are known as solar active

regions. They are identified by a variety of features such as sunspots,
pores, filaments and prominences, plages, and enhanced Xray emission.
A filament is a cloud of material above the solar surface which is
cooler and slightly denser than its surroundings and appears as a dark
stringlike feature on the solar disk when viewed in any strong emission
line; most of the existing data is in H-a, the easiest fine in which to
observe these features (Zirin, 1988). Filaments generally occur in the
chromosphere along magnetic neutral lines where the local magnetic field
is parallel to the solar surface between two regions of opposing flux.
They may also form from material ejected from solar flares or condense
out of coronal material as coronal rain (Zirin, 1988).
A prominence is identical to a filament with the exception of the
way it is observed. Prominences are filaments that are observed on the

limb of the solar disk and appear as bright loops protruding from the
edge of the Sun. Even though the material contained in these structures
is cooler than the local environment, they appear bright when viewed
against the blackness of space at the solar limb.
A plage (the French word for beach) is an area in or near an active
region that appears bright in H-a or other strong emission line compared
to the background quiet sun. Plage regions show a moderate amount of
emerging magnetic flux and indicate a region of gas with a higher density
than the surrounding atmosphere and a hotter overlying corona (Gibson,
1973). If the magnetic flux is increasing, the local field will strengthen
and eventually inhibit the convective flow of energy from below the
surface. The result will be a localized region of material that is colder
than its surroundings and a strong magnetic field (a pore or sunspot)
indicating the birth of an active region. Decreasing magnetic flux in a
plage, the decay phase of an active region, shows weakening magnetic
fields and increasing convective turbulence, eventually leading to total
disintegration of an active region. Plages, therefore, define both the
spatial and temporal extent of a solar active region.
Sunspots, cited by Galileo in the 1600s as evidence to the Catholic
church that the Sun is not a perfect, unchanging celestial object, appear
as dark blemishes of various sizes on the solar disk between about 30
north and 30 south solar latitude (Snow, 1987; Zirin, 1988; Hartman,
1991). Sunspots are regions in the photosphere of intense magnetic
fields, as mentioned above, which have successfully suppressed the
convective transport of heat from below the solar surface. The plasma
within a sunspot is thereby cooled from the ambient 5700 K of the

photosphere to approximately 4000 K,
causing the spot to appear dark (Zirin,
1988; Hartman, 1991). The overall
structure of a single sunspot, in white light
(Figure 2.2), is generalized into two
features, the umbra and penumbra. The
umbra is the central dark core of a sunspot
where the strong vertical magnetic field is
concentrated. The penumbra is the lighter,
grayish boundary that surrounds the
umbra and contains more horizontal
magnetic flux tubes.
Pores are features which generally signal the initial formation of a
sunspot in white light and appear as the dark umbral centers of sunspots,
but without a penumbra. The exact process which forms a penumbra
around a pore to define a sunspot is unknown. However, sunspots have
been observed to form from the coalescence of multiple pores, and by an
increase in the magnetic field strength of a pore.
Once a sunspot group or active region forms on the solar surface, it
is classified according to its magnetic configuration. In 1919, a group at
the Mt. Wilson Observatory, lead by George Ellery Hale, introduced the
Mt. Wilson magnetic classification for active regions which is still in use
today (Zirin, 1988):
Figure 2.2 A white light
image of a typical sunspot,
illustrating the umbra and
penumbra features of a sunspot.

a: A single dominant spot, usually connected to a plage of
opposite magnetic polarity.
(3: A pair of dominant spots of opposite polarity.
y: Complex groups with irregular distribution of polarities.
Py: Bipolar groups with no marked north-south inversion line.
5: Umbrae of opposite polarity in a single penumbra.
In addition, p and f suffixes are used when the preceding or following
spot, respectively, is dominant (Zirin, 1988).
The Solar Cycle
The Sun, for reasons unknown, reverses its magnetic field every
eleven years, described as the solar cycle, which affects the number,
position, and polarity of sunspots. Shortly after the beginning of a cycle,
in a simplified scenario, the number of sunspots is at a maximum, and
the majority of them form at high latitudes around 30 north and south
(Hartman, 1991). As spots emerge, they do so as magnetically opposing
pairs with the preceding spot (pspot), in terms of the direction of solar
rotation, having the same polarity as the pole which it is nearest, and the
following spot (fspot) having opposite polarity (Zirin, 1988; Hartman,
1991). For example, if the solar north polar region is positive polarity for
a particular cycle, then the leading sunspots of active regions in the

northern hemisphere will tend to have positive polarity for that cycle as
well. As the cycle progresses, sunspots gradually begin forming closer
and closer to the solar equator with
11 years time
Figure 2.3 A sketch illustrating the
variation in sunspot latitude of
formation with the 11 year solar cycle.
less frequency and complexity until
they eventually cluster around 10
latitude at the end of the eleven
years. At this time the cycle starts
over again except the Suns polarity
is now opposite from the last cycle.
Figure 2.3 is a sketch of this process
which demonstrates the variation in
latitude of sunspot location with the
passage of one solar cycle.
Proper Motion of Sunspots
To make the study of active regions even more complicated,
sunspots move relative to each other. As sunspots and new flux regions
emerge and exist on the solar surface, they tend to do so in the preferred
configuration of having the leading spot(s) the same magnetic polarity as
the polarity of the pole they are nearest (Zirin, 1988; Hartman, 1991).
Based on personal observations, the motion is most noticeable when a
region emerges reversed from the preferred configuration. From the
onset of plage and spot formation, the individual flux regions begin to

move (longitudinally for the most part) relative to one another in order to
attain their preferred magnetic configuration. Even if regions emerge
with their nominal orientation, relative proper motion between spots still
occurs (Mazzucconi, Coveri, & Godoli 1989; Howard, 1991). In the study
conducted by Mazzucconi, Coveri, & Godoli (1989) on over 400 sunspot
groups, their results showed that only 4% of the regions examined did not
show systematic proper motions and that, on average, p-spots move
forward in longitude (with respect to the direction of solar rotation) about
0.5 per day and f-spots move backward about 0.5 per day.
Presently, the most widely accepted theory of sunspot proper
motion is centered on the concept of looped, kinked, knotted, or twisted
magnetic flux ropes rising from below the photosphere due to magnetic
buoyancy (Snow, 1987; Zirin, 1988; Mazzucconi, Coveri, & Godoli, 1990;
Tanaka, 1991; Lites et al., 1993; Skumanich et al., 1995). It is presumed
that such twisted knots are formed in the convection zone of the Sun by
forces analogous to twisting a rubber band (Tanaka, 1991). As the
deformed flux rope penetrates the photospheric surface, sunspots form at
the locations where the rope exits and returns to the solar surface due to
the localized strong magnetic field strength. Continual ascent of the flux
tube will then expose lower and lower cross-sections of the knot,
producing apparent photospheric motion (Zirin, 1988; Tanaka, 1991;
Skumanich et al., 1995).
Delta (5) Sunspots
The most dynamic and intriguing types of sunspots to study are

those which exhibit a delta (8) configuration, i.e. umbrae of opposite
polarities within the same penumbra (Zirin, 1988; Tanaka, 1991;
Skumanich et al., 1995). A 8spot can form in one of two ways, either by
the emergence of new flux of opposite polarity within an already existing
spot, or by the collision of two spots of opposite polarity resulting from
photospheric motion (Zirin, 1988; Tanaka, 1991; Gaizauskas, Harvey, &
Proulx, 1994). Typically, a low-lying filament will form between the
umbrae which, observationally, reveals the path of the neutral line (Zirin,
1988; Hongqi, 1993; Skumanich et al., 1995). As the two spots begin to
interact, their geometric circular form often becomes distorted due to the
compression of the penumbral fields (Zirin & Wang, 1993). The
penumbral fibrils, which can be thought of as proxy tracers of the
horizontal magnetic fields (Hongqi, 1993; Gaizauskas, Harvey, & Proulx,
1994), no longer show a radial structure; they have a cyclonic appearance
which indicates a twist in the overall magnetic field of one, or both, of the
regions, and a high degree of shear across the neutral line (Zirin, 1988).
As the interaction proceeds, the fibrils begin to overlap indicating a direct
magnetic link across the neutral line and the region is officially in a 8
configuration (Gaizauskas, Harvey, & Proulx, 1994).
The shear angle in an active region is defined as the angle between
neutral line and the overlying horizontal field as measured from a line
perpendicular to the neutral line, i.e. if the neutral line and the
horizontal field form an angle of 90, the shear angle is zero (Ambastha,
Hagyard, & West, 1993; Chen et al., 1994; Zhang, 1994). A high shear
angle indicates that the local magnetic gradient is steep and has been
used in the past as a criterion for the prediction of solar flares: the larger

the shear angle, the steeper the magnetic gradient, and the higher the
probability for flare activity. In a study of twenty large flares, Chen et al.
(1994) determined an average shear angle of 56 for flaring neutral lines,
suggesting a relatively strong magnetic gradient which could be used an
indicator for potentially large flares. However, in a more recent study,
Smith et al. (1996) examined 57 active-region-days, covering 35 different
active regions, and concluded that the shear angle is not a viable
predictor of solar flares.
The Physics of Modeling Sunspots
Determination of an active regions magnetic field strength is
accomplished by analyzing spectral absorption lines which exhibit the
Zeeman effect. The Zeeman effect is the splitting of spectral lines caused
by an external magnetic field which interacts with the spin-orbitals of
the atoms in the plasma by removing the degeneracy in the orbital (mj)
quantum number (Bransden & Joachain, 1983). The energy difference
between the two split (a) lines is given by the relation
AE = gpB Bz mj (2.1)
where pb is the Bohr magneton, Bz is the z component of the magnetic
field, m, is the component of the angular momentum parallel to the
magnetic field, and g is the Lande factor given by (Solanki, 1993;
Bransden & Joachain, 1983)
j(j + l)+s(s + l)~ 1(1 + 1)
2j(j + l)

Further information about the strength and direction of the
observed magnetic field can be acquired by examining the polarization of
the light of the unsplit (71) line and the a lines (see Figure 2.4). Polarized
light is generally described by the four Stokes parameters I, Q, U, and V,
where I is the total intensity (the sum of the polarized and the
unpolarized fractions of light), and (Solanki, 1993; Hecht, 1987)
Q Ilin(x 0) Ilin(x ~ %) (2.3 a)
U = Iii(x = ^)-Ii1(x = 3^) (2.3 b)
V = Icirc(right)-Icirc(left)- (2.3 c)
Here Iun(x) is the linearly polarized light whose electric vector makes an
angle x to some reference direction defined by the measuring instrument,
and Lire refers to circularly polarized light
(Solanki, 1993). Depending on the
orientation of the observation of the
magnetic field (that is, whether the line of
sight is parallel or perpendicular to the local
magnetic field), the 71 and a lines will show
various states of polarization which are then
used to define the Stokes parameters above.
Table 2.1 shows the polarization results for
each of the lines with the extreme cases of
strictly longitudinal or strictly transverse
observations. In the case of completely
Figure 2.4 An edge
enhanced image of the Fe I
6302 A line displaying the
magnetic splitting of a
spectral line into its n and a

polarized light, the Stokes parameters are no longer independent of each
other and the Stokes vector can be adequately described by only three of
the parameters: I2 = Q2 + U2 + V2 (Hecht, 1987; Solanki, 1993).
Split Spectral Line Component Longitudinal B Field Observation Transverse B Field Observation
71 plane polarized absent
a + left hand polarized plane polarized
a - right hand polarized plane polarized
Table 2.1 Polarization of Zeeman split spectral lines observed in a
magnetic field parallel and perpendicular to the line of sight (Zirin, 1988;
Bransden & Joachain, 1983).
Once the strength and direction of the active region magnetic field
is determined by Stokes analysis, this information is then incorporated
into magnetohydrodynamic (MHD) models which describe the behavior of
a plasma, as a fluid, in a magnetic field. The complete set of fluid
equations for a two species, ions and electrons, plasma are (Chen, 1984;
Zirin 1988):
Charge and current densities
ct= + neqe
(2.4 a)
j= n^iVi +neqeve =o(E + vxB)
(2.4 b)

Maxwells equations
VxE = -B
V B = 0
e0V-E = + neqt
VxB=niqivi+neqev0+eo E
Continuity equations
Mvrv)vj =(ijnj(E+vjxB)-vpj
- + V -(iijvj) = 0
J = i,e
Equation of state
pi= J= i,e
y = gas constant
(2.5 a)
(2.5 b)
(2.5 c)
(2.5 d)
(2.6 a)
(2.6 b)
Maxwells equations describe the relations between the electric field E,
the magnetic field B, the charge density, and the current j; the
hydrodynamic equations interrelate pressure, density, temperature, and
flow (Zirin, 1988).

Magnetic Structure of Sunspots
Large stable sunspots are typically characterized by a round
umbra-penumbra form (Zirin, 1988). In the central dark umbra there is
a strong vertically oriented magnetic field, typically on the order of 3,000
gauss (Zirin, 1988; Zirin & Wang, 1992; Gaizauskas, Harvey, & Proulx,
1994; Lites et al., 1993). The temperature inside the umbra is
approximately 4,000 K (Zirin, 1988; Hartman, 1991), causing it to appear
darker than the surrounding photosphere, and the density is roughly
equal to that of the photosphere (~ 0.4 g/cm3) in order to maintain
hydrostatic equilibrium (Zirin, 1988). The low temperature, and
therefore the dark appearance, is due to radiative losses caused by
suppression of the vertically directed convective thermal transport
resulting from the magnetic pressure produced by the strong field of the
emerging flux (Gaizauskas, Harvey, & Proulx, 1994). This effect can best
be understood by considering the pressure balance between the spot and
the photosphere (Zirin, 1988; Solanki, 1993):
_+n k T =n k T (2.8
gJt+nspKBisp nphKBiph
where the subscripts sp and ph stand for values within the spot and the
photosphere, respectively, kB is the Boltzmann constant, T is the
temperature, n is the number density, and B is the magnetic field

Surrounding the umbra with radial starburst appearance is the
penumbra. The most prominent feature of the penumbral magnetic field
is the presence of azimuthally narrow fibrils, or spines, of enhanced
magnetic field which are generally horizontal to the solar surface (Lites
et al., 1993). Detailed high resolution images of sunspot penumbrae
show that the penumbral fibrils are filament like structures azimuthally
separated by relatively small distances. The lighter grayish appearance
is due to this separation which allows the hotter and brighter background
photosphere to protrude through the gaps between the fibrils. According
to Wentzel (1992), the strong umbral magnetic field can be described as a
bundle of flux ropes extending vertically from the solar surface, and the
horizontal penumbral fibrils as individual flux ropes that have fallen
from the umbral bundle. He argues that fallen flux tubes are
energetically favored by the shape of the sunspots large-scale magnetic
field and that the boundaries of the penumbra defined by fallen flux
tubes tend to coincide with the boundaries specified by the (competing)
return flux model of spot fields (Wentzel, 1992). The field strength B in
this region, as determined by Stokes analysis, is typically in the range
800 G < B < 1600 G (Solanki, Montavon, & Livingston, 1994) and
accounts for a significant fraction of the total magnetic flux of a sunspot
(Lites et al., 1993).
Above the sunspot, in the lower chromosphere, the pressure and
density decrease with height; the scale height is roughly 200 km (Zirin,
1988; Judge, 1994). As the vertical magnetic field of the sunspot
penetrates the chromosphere, it must compensate for the change in
pressure of the external environment by increasing its volume to

maintain hydrostatic equilibrium. The result is a magnetic canopy
which extends well beyond the outer boundary of the penumbra (Lites et
al., 1993; Skumanich et al., 1995) to a distance of approximately 2xl04
km, and to a height of about 350 km above the photosphere (Solanki,
Montavon, & Livingston, 1994). This structure has been successfully
observed in the solar atmosphere, most notably by the Advanced Stokes
Polarimeter group at HAO/NCAR, and is identified by regions of low
magnetic field strength (Skumanich et al., 1995) at the height of
formation of the Fe I 630 nm line (Lites et al., 1993).
Once in the chromosphere, the bulk of the sunspot magnetic field
will be connected to one or several other flux tubes of opposite polarity
from surrounding sunspots and plages. In large complex active regions,
however, the outer boundary of the chromospheric field will even extend
into the corona where it will become part of the large-scale magnetic
structure of the Sun in the form of a coronal loop (Gosling, 1993). Values
of the coronal magnetic field strength above sunspots have been
determined by Gary et al. (1993) by analysis of microwave radiation
between 3 and 15 Ghz, attributed to gyroresonance radiation at the
second and third harmonics of the gyrofrequency of high-energy
electrons. The values they obtained were on the order of ~ 1000 1500
gauss. However, a small simple static two spot dipole region with the p-
spot leading the f-spot will generally not extend into the corona. In this
type of simplified generic scenario, the magnetic structure is analogous to
a common bar magnet being placed just below the Suns surface with the
lines of force coming off of one end of the magnet, arching up through the
chromosphere, and returning to the other end of the magnet.

Solar Flares
Solar flares are intense, abrupt releases of energy which occur in
areas where the magnetic field is changing because of flux emergence or
sunspot motion (Zirin, 1988; Tanaka, 1991; Hongqi, 1993). They occur in
the magnetic field lines across an active regions neutral line where
strong shear exists, as mentioned above, and result in a lower energy
configuration for the local magnetic field (Zirin, 1988; Chen et al., 1994).
In H-a, at 6563 A (the most common spectral line used to observe the
lower chromosphere and photosphere), flares appear as intense
brightenings, in what are termed ribbons and kernels, and generally
reveal changes in the filament structure of the region and post flare
The solar flare itself is analogous to twisting a rubber band until it
breaks, where the rubber band represents the highly sheared magnetic
field lines in an active region. The stress in the magnetic field steadily
increases, due to new flux or motion, until the energy configuration of the
field is no longer favorable and the field lines reorganize in a flare.
There are several possibilities for the energy release mechanism in flares,
none of which is known to be absolutely correct every time. One theory
suggests that the field lines across the neutral line suddenly shorten and
in doing so release energy (Zirin, 1988). Another leading theory claims
that magnetic reconnection of the field fines is the energy source of flares
(Demoufin et al., 1993), resulting from the neutral fine filament lowering
the conductivity and shortening the diffusion time of the region (Zirin,
1988). More recently, Leka et al. (1993) have proposed that an instability

driven by vertical currents is responsible for the onset of a flare and
particle acceleration. In any case, the exact mechanism that triggers a
solar flare is unknown.
Solar flares have been classified in many ways, but probably the
most common convention is defined by the flux of soft X-rays emitted by
the Sun. In this classification scheme, decadal increments of flux,
starting with ICh* Watts/m2, are associated with the letters A, B, C, M,
and X followed by the logarithmic level associated with the particular
decade. For example, if the flux rises to a level of 2.5xlO-A then the
associated flare would be classified as a C2.5.
At the onset of a flare, electrons and ions are accelerated along the
magnetic field fines into the photosphere (Zirin, 1988). In large flares,
electrons can be accelerated to energies of more than 100 MeV and ions to
energies of up to several GeV (Droge et al., 1989). As the particles impact
the higher density plasma of the photosphere, they are rapidly
decelerated producing the observed X-ray spectrum through the
bremsstrahlung process. The impact sites of these particle beams are
known as flare kernels, or foot points (Canfield et al., 1990; Zirin & Tang,
1990; Winglee et al., 1991; De La Beaujardiere, Kiplinger, & Canfield,
1992). When viewed in H-a, the flare kernels appear as intensely
brightened regions on both sides of the neutral fine due to the heating
caused by the injection of the high-energy particles. As the flare
progresses along the neutral fine in a cascade effect, the flare kernels
follow, forming two bright ribbons parallel to the neutral fine (Zirin &
Tang, 1990; De La Beaujardiere, Canfield, & Leka, 1993; Hongqi, 1993).
During this impulsive phase of a flare, the overlying neutral fine filament

is disrupted, sometimes by energetically erupting from the solar surface
which can be easily observed in the blue wing of H-a (Zirin, 1988;
Canfield et al., 1990).
In the final stages of a flare, the magnetic field lines have been
reorganized into a lower energy configuration, relieving some of the
stress of the region. At this point, H-a emission is decreasing and the
formation of post flare loops can be observed; post flare loops appear in
H-a as filament like structures curving over the neutral line. They are,
however, the active regions field lines draining coronal material which
was acquired during the flare in field lines that opened to the corona
(Zirin & Tang, 1990). The hot coronal material confined to the
chromospheric magnetic field of the region strongly emits in H-a and
exposes the geometry of the field lines.
Time profiles of a flare can vary anywhere from a few seconds to
hours. Long duration events, lasting more than about an hour, typically
occur with fields which are connected to the large-scale magnetic
structure of the Sun and loops that extend well into the corona (Gosling,
1993). These events are associated with coronal mass ejections (CMEs),
interplanetary proton events, and magnetic storms (Parks, 1991; Zirin &
Tang, 1990; Gosling, 1993). Long duration flares also show a hardening
of the soft X-ray spectrum, whereas impulsive events show a powerlaw
spectrum (Droge et al., 1989). Impulsive flares tend to be more localized
than long duration events and occur with ribbons in close proximity to
the neutral line. This could imply that impulsive flares represent
relatively small localized areas of strong shear based on the conclusions
of Chen et al. (1994). In their study of solar flares, Chen et al. (1994)

determined that if the initial brightenings are close to the neutral line,
then the neutral line tends to have higher shear, and if the initial
brightenings are away from the neutral line, the neutral line tends to be
less sheared.

The digital images and data used in this study were obtained from
several ground-based observatories and one Earth orbiting satellite. The
ground-based observatories include the Vacuum Telescope at the Kitt
Peak National Solar Observatory near Tucson, Arizona, and three of the
five facilities of the SOONSPOT observatory network. The space-borne
data is from the Japanese Soft Xray Telescope on board the Yohkoh
satellite. All of the data and images were reduced using the Interactive
Data Language (IDL) computer program.
Currently, no literature exists which describes the physical aspects
of the SOONSPOT telescopes. However, Dr. Alan Kiplinger, coordinator
of the SOONSPOT observatories, has generously provided basic
information about these facilities.
The telescopes used at the SOONSPOT locations are 10 inch
refractor heliostats which acquire both white light and H-a large scale
images of solar active regions, and full disk H-a images. The large scale
field of view is approximately 6x6 arcminutes with an average pixel size,

between observatories, of 0.7 arcseconds. The white light and H-a full
disk images are taken with a cadence of thirty minutes, and the large
scale H-a images are acquired with a cadence of either five minutes or
thirty seconds, depending on flare activity.
The images themselves are digitized from video and then stored on
8-mm Exabyte data tapes as extended FITS (Flexible Image Transport
System) files. Each image is stored as a 512 x 512 byte array with an
accompanying 72 element string array header which describes specific
aspects of the image. The data contained in the image headers includes
information such as the observation date and time, observatory name,
type of observation (H-a or white light), type of object observed (full disk
or large scale), pixel resolution, and image coordinates of a reference
pixel. The image headers also include information about the image
reference pixel which is related to the region of the Sun that is being
observed. These items are listed and defined in Table 3.1 for later
Item Description
Position angle The angle from the northern most extremity of the solar disk to the reference pixel.
Radius vector The distance, in solar disk radii, from the center of the solar disk to the reference pixel.
Heliocentric longitude and latitude The longitude and latitude of the reference pixel in the Sun-centered coordinate frame
Solar B angle The tilt of the Suns rotation axis toward or away from the line of sight with the Earth, measured positive to the north.
Solar P angle The east-west tilt of the Suns rotation axis as viewed from Earth, measured positive to the east.
Table 3.1 Definitions of SOONSPOT reference pixel items, related to an observed
region of the Sun, which are stored in the image headers.

The X-ray images used here are from the Soft X-ray Telescope
(SXT) on board the Japanese Yohkoh satellite which was placed in
operation in 1991. The main scientific objectives of the SXT instrument
are to study the geometry of the coronal magnetic field topology, the
temperature and density characteristics of Xray emitting plasma, the
spatial and temporal characteristics of flares, the transport of energetic
particles, the locations of energy release and particle acceleration, and
the presence of waves or other magnetic field disturbances associated
with sprays, filament eruptions, and coronal transients (Tsuneta et al.,
The SXT is a glancing incidence telescope which utilizes
hyperboloids of revolution for the optical surfaces of the incident mirrors
to focus the X-rays on a flat focal plane with moderate wide-field angular
resolution (Tsuneta et al., 1991). The X-rays are brought to focus
through a 50 mm aperture on to a 1024 x 1024 virtual phase charge
coupled device (CCD) detector at a focal length of 1.54 m (Tsuneta et al.,
1991). X-ray images in the range 0.25 to 4.0 keV are acquired by the use
of filters centered on specific wavelengths: Al 1265 A, Al/Mg/Mn
composite 2930 A/2070 A/562 A, Be 119 pm, Al 11.6 pm, and Mg 2.52 pm
(Tsuneta et al., 1991). By computing intensity ratios of various filter
combinations, solar plasma temperatures can be estimated with an
uncertainty of 0.1 in logioT for isothermal sources (Tsuneta et al., 1991).
Data from the SXT is transferred as 1024 x 1024, 512 x 512, or
256 x 256 images, depending on whether 1 x 1, 2 x 2, or 4 x 4 on-chip

summation is used (Tsuneta et al., 1991), with image resolutions of 2.45,
4.91, and 9.81 arcseconds per pixel, respectively (Morrison, 1994b). The
SXT image data are stored as 8bit number arrays which have lower
order accuracy than the 12-bit analog to digital converter employed by
the CCD detector. The full accuracy of the data is restored by a
decompression scheme which is described by Tsuneta et al. (1991) and
Morrison (1994c). The processing and manipulation of the Yohkoh data
is made effortless by an extensive IDL software library which handles all
of the image decompression, spacecraft pointing history information, and
image characteristics.
Kitt Peak
Photospheric longitudinal magnetograms were provided by Kitt
Peak National Solar Observatory with the Vacuum Telescope. The
Vacuum Telescope is a four mirror coelostat with the light path contained
in a vacuum (Livingston et al., 1976), and adaptable to a variety of
instruments. Specifically, photospheric magnetograms are obtained with
the spectromagnetograph which images the circular polarization (Stokes
V) signal of the Fe I 8688 A spectral line (Jones et al., 1992). Full disk
magnetogram images of the Sun are generated by stepping the slit of the
spectrograph across the Sun, and then placing columns of magnetic
information from each image side by side to form an image of the Sun
(i.e., the magnetic information along the magnetograph slit from each
spectral image forms one column of the full disk image). The full
scanning process takes about an hour from beginning to end.

The resulting images from Kitt Peak are 1788 x 1788 16bit data
arrays with lxl arcsecond pixel resolution. The images are stored in
standard FITS format with 60 element string array headers which
describe the aspects of the image and observation.

The SOONSPOT white light images of NOAA AR 7978 were
analyzed to determine sunspot proper motions for this region, and to
attempt to relate penumbra formation with umbral size. The importance
of this type of analysis is that it yields some sense of the three-
dimensional structure of the rising magnetic flux tubes. By tracking
sunspot sizes and motions, and assuming that the emerging magnetic
flux is indeed in the form of a tube, one can extrapolate a reasonable
geometry of active region flux tubes, using the sunspots as photospheric
The sections of this chapter address the proper motions of the
dominant sunspots of AR 7978, and discuss the techniques used to
determine each sunspots heliocentric location. In particular, the proper
motions of the two umbrae in the 5configuration of AR 7978 are
examined, along with the motions of a small third spot which may have
played a roll in the formation and disintegration of the 8-spot. In the
second half of this chapter, I discuss my investigation of penumbra
formation as related to umbral size.

Sunspot Proper Motions
Due to variable seeing and the large area of the Sun which is
covered by the SOONSPOT white light data, exact locations of sunspot
features are difficult to discern on many of the images. For this reason,
the accuracy of the relative sunspot positions in the images is not as
desirable as one would expect and leads to an increase of the error for
calculated heliocentric coordinates of each spot. However, for the
technique described below, I believe I have minimized this source of error
and that my selections of umbral centers for each sunspot on the images
is to within two pixels in both the x and y dimensions.
For reference, Figure 4.1a shows AR 7978 in a well developed stage
of its formation, and the labeling scheme which is used in this analysis.
Figure 4.1b is a photospheric longitudinal magnetogram of the region
taken approximately one and a half hours before the white light image in
Figure 4.1a. The dominant sunspots are labeled A, B, C, and D, the
opposite polarity umbra of the 8spot is labeled A5, and the small spot
labeled AD is the one which may be associated with the formation and
disintegration of the 5-spot. The 5-spot itself is in the box of Figure 4. la
and is easily distinguishable in the center of the magnetogram image.
For the purposes of this study, the white regions of the magnetogram are
defined as positive polarity, the black regions as negative polarity, and
the gray as neutral.

Figure 4.1 NOAA AR 7978 in a well developed stage of its formation, with the 8-
sunspot indicated by the black box (a), and a photospheric longitudinal magnetogram
of the region taken between 14:47 and 15:50 UT 8 July 1996 (b). The dominant spots
are labeled AD, the opposite polarity umbra of the 8spot is labeled A8, and AD is a
spot which may have had a significant effect on the 8-spot.______________________

Analysis Technique
Computing Solar Coordinates. To determine the radius vector,
position angle, and heliocentric longitude and latitude for any given pixel
in a SOONSPOT image requires several steps. The first step is to
determine the distance, in pixels, from the center of the solar disk to the
image reference pixel which is recorded in each image header. The
information necessary to determine this value is the pixel resolution of
the image (arcseconds), the angular radius of the solar disk (arcseconds),
and the radius vector and position angle of the reference pixel. The
angular radius of the solar disk for the observation dates of 7-9 July 1996
were obtained from The Astronomical Almanac For The Year 1996 (1995)
which lists the angular diameter of the Sun for each day of 1996.
Once this information is acquired, the radius of the solar disk, in
pixels, is calculated by dividing the angular radius of the solar disk by
the pixel resolution of the image. For example, if the pixel resolution is
0.7 arcsec/pixel, and the radius of the solar disk is 950 arcsec, then the
radius of the solar disk, in pixels, is
qca arcsec
^ solar radius pixels (4-1)
n 1 arcsec solar radius
U* pixel
This value is calculated for both the x and y dimensions of the image and
then an average value is taken, since the pixel resolution of the
SOONSPOT images varies slightly between the two dimensions. This
leads to, at most, a 10% error by comparison to the original image
resolution. The distance of the reference pixel from the center of the

solar disk, in pixels, is then computed by multiplying the solar disk
radius, in pixels, by the radius vector from the image header.
Next, the reference pixel is placed in a solar disk (x, y) coordinate
system (see Figure 4.2) which has its origin at the solar disk center, with
celestial west corresponding to the positive x direction (the right side of
the SOONSPOT images) and celestial north corresponding to the positive
y direction (the top of the SOONSPOT images). The reference pixel
position angle is then converted to a unit circle position angle (PAu),
where 0 corresponds to celestial west, 90 corresponds to celestial north,
and so on. The solar disk (x, y) coordinates are then determined, using
the radius vector in pixels (RVp), and the sine and cosine trigonometric
Figure 4.2 Schematic diagram of the method used to determine heliocentric
coordinates for individual SOONSPOT image pixels.__________________________________

X coordinate = RVP Cos(PA
(4.2 a)
Ycoordinate = RVP Sin(PAu) (4.2 b)
Once the solar disk (x, y) coordinates of the reference pixel have
been determined, the coordinates of any other pixel in the image can be
obtained by simply adding or subtracting their x and y offsets from the
reference pixel to the solar disk (x, y) coordinates of the reference pixel.
For example, if the image reference pixel is (243, 263) and its solar disk
coordinates are (400, 350), then an image pixel at (273, 203) would have
solar disk coordinates of
(400 + [243 273], 350 + [263 203]) = (370, 290). (4.3)
Once the solar disk (x, y) coordinates of an individual pixel in an
image are known, the position angle and radius vector of that pixel can
then be calculated by the reverse process of that described above. From
the position angle and radius vector of a pixel, the heliocentric longitude
and latitude of the point are determined from the following equations:
Sinfp+p^-^ (4.4 a)
Sin(B) = Sin(B0)Cos(p) + Cos(B0)Sin(p)Cos(P -0) (4.4 b)
Cos(B) Sin(L L0) = Sin( p) Sin(P 0) (4.4 c)
where p is the heliocentric angular distance on the solar surface from the

center of the Suns disk, pi is the radius vector, 0 is the position angle, S
is the semidiameter of the Sun (radius of the solar disk), P is the solar P
angle, B is the solar B angle, B0 is the heliocentric latitude which is
measured positive to the north, and the quantity L L0 is the heliocentric
longitude as measured positive to the west of the central meridian (The
Astronomical Almanac For The Year 1996, 1995).
Determining Sunspot Centers. To determine which point of a
sunspot to consider as a representation of its position, intensity contours
of the white light images were used to select a pixel in the darkest
umbral region of each spot. To accomplish this, an IDL procedure was
written which displays both the image and a plot of its contours, with the
maximum intensity contour set at 75% of the minimum-maximum
intensity range of the image and a limit of three contours. This
technique of displaying the contours was chosen for its ability to display
dark umbral cores, even in images with bad seeing. Figure 4.3 illustrates
the utility of this procedure with a comparison between an image and its
contours with good seeing and one with bad seeing.
The representative position pixel is then chosen by the user by
centering the cursor in the darkest umbral contour and clicking the
mouse button. User selection of the position pixel was chosen over
computerized selection because of the variable odd sunspot geometries,
intensities, and seeing which would have been difficult to account for in a
program. However, by using this method of contours, selecting the center
of the darkest umbral core with the computer cursor is fairly accurate
because of the small regions identified by the contours. In an image with

a sunspot which measures 35 pixels in diameter, and with bad seeing, an
error of two pixels in the positioning of the cursor gives an estimated
error of 6% in the calculations of the sunspot position, or 0.1 in
heliocentric longitude and latitude. Of course, images with better seeing
have a wider range of intensities, and therefore the lower intensity values
are better defined in the contours, which reduces this source of error.
Proper Motions of the Major Sunspots
The proper motions of the dominant sunspots of AR 7978 were
measured relative to sunspot A which is visually a natural center to the

entire active region. The relative heliocentric positions of sunspots B, C,
and D, in relation to sunspot A, are illustrated in Figure 4.4. The top plot
in Figure 4.4 shows the rotation angle from solar north of spot A for each
of the dominant sunspots. The bottom plot shows the heliocentric
longitudinal separation between each of the dominant spots and spot A,
with positive values measured to the west.
One interesting aspect of these plots to note is that the most rapid
motion of spots B and D occurred during the first day of observations
which is indicated by the slope of the points in the longitudinal motion
plot. This suggests that the initial motion of the sunspots my be due, in
part, to the rising flux tube responding to conditions in the
chromospheric environment, as well as the apparent motion produced by
lower and lower cross-sections of the flux tube being exposed as it rises
through the photosphere. This would indicate that chromospheric
conditions, such as plasma flows, the ambient magnetic field, and
pressure gradients, may play an important role in the initial
configuration of an active regions magnetic structure.
A second aspect to notice in the plots of Figure 4.4 is the
approximate 20 clockwise rotation the entire active region underwent
throughout the observation period. This slow rotation indicates that
there was a slight twist to the macroscopic magnetic field of the active
region which was exposed as it ascended through the photosphere. The
most probable cause for the twist would be subphotospheric motions
twisting the flux tube before it penetrates through the photosphere.

Rotational Motion of Dominant Spots
Longitudinal Motions of Dominant Spots
Figure 4.4 Proper motion plots of the dominant spots of AR, 7978,
with respect to sunspot A. The upper plot shows the rotation angle of
each spots position from solar north of spot A. The lower plot shows the
heliocentric longitudinal distance (degrees) of each spot from spot A.

The lack of large variation in position between the dominant
sunspots of this active region indicates that the region, as a whole,
emerged in its preferred orientation, with the leading spots the same
polarity as the polarity of the southern hemisphere in which it resided.
This conclusion is supported by the time series photospheric
magnetograms in Figure 4.5 which shows the magnetic polarity of the
region from 48 July 1996. As illustrated by these images, the positive
(white) polarity maintains its leading position from before the emergence
of the first sunspot, through the complete formation of the active region.
17:10:36 18:13:01 14:14:22 15:09:00 14:26:37 15:25:58
\. A- < tft:
07/04/96 07/05/96 .. 07/08/96
14:39:52 15:34:32 14:46:45 15:50:50 t
I Celestial
^ 07/07/96 07/08/96
Figure 4.5 A daily time series of Kitt Peak photospheric longitudinal
magnetograms of AR 7978. The date of observation is at the bottom of each image,
and the UT start and stop times of the full disk magnetograms is at the top of each

These results of the proper motions of AR, 7978 presented here are
similar to those found in a study of 408 sunspot groups by Mazzucconi,
Coveri, & Godoli (1989). In their investigation, it was found that only 4%
of the examined groups showed no systematic proper motions at the
beginning of their formation (Mazzucconi, Coveri, & Godoli, 1989). It was
further concluded that, on average, the p-components move forward
(westward) about 1 in longitude per day, and the f-components move
backward in longitude approximately 0.5 per day (Mazzucconi, Coveri, &
Godoli, 1989).
Proper Motion of the Delta-Snot
As the opposite polarity umbra which defined the 5spot, A8,
emerged on 8 July, its location was approximately 0.3 west and south of
spot A, where it remained for the rest of the day (see Figure 4.6).
However, shortly after A5 appeared, spot AD emerged 0.5 to the north
and west of sunspot A and appeared to have a relevant magnetic
connection to AS through filamentary structures (see Figure 4.7). Spot
AD immediately began moving (Figure 4.6), relative to A, once it
emerged. In the first four hours of observations, from 6:01 to 10:01 UT,
AD moved 0.5 in both longitude and latitude. By the end of observations
on 8 July, sunspot AD was 1.25 west and south of spot A and was
heading toward sunspot D.
During the data gap from 16:47 on 8 July to 4:51 on 9 July, both
sunspots, AS and AD, showed a significant clockwise rotation about

Rotational Motion of 5-Related Spots
Longitudinal Motions of 6-Related Spots
08Jul96 09-Jul96
Observation Time (Axis Start Time 07-Jul-96 05:00:00 )
Figure 4.6 Proper motion plots of spots A6 and AD, with respect to
sunspot A. The upper plot shows the rotation angle of each spots
position from solar north of spot A. The lower plot shows the
heliocentric longitudinal distance (degrees) of each spot from spot A.

* y
AS A <5
0/07/96 06:01:32 0/07/96 10:01:33
* ' M ^ * 1 W fi4 -
A 6 1
0/07/96 16:47:33 9/07/96 04:51:33
C f * f
\ ^ AS 1
9/07/96 06:51:33 9/07/96 10:24:33
Figure 4.7 An abbreviated time series proper motions of sunspots A8 and AD. of the life-cycle of the 5-spot, illustrating the
sunspot A. Spot A8 rotated approximately 70 and AD rotated about 20,
with A8 maintaining its relative distance from spot A. This similar
concurrent motion of these two sunspots is a second indication that they
were directly magnetically connected, perhaps as the footpoints of a
single, low lying flux tube.
The large, and somewhat quick, rotation of spot AS about spot A
implies that there was an intertwining of the flux tubes defining these

two sunspots (see Figure 4.8). As the region
developed, the single twist between these
two flux tubes may have ascended through
the photosphere, appearing as a large
rotation of spot A8 about spot A.
For the remainder of the last
observation day, spot AD continued to move
toward spot D and separate from AS. In
Figure 4.7, the image of 6:51 UT on 9 July
shows nearly the last observation in which
spots AS and AD have intermixed penumbral
structure connecting them. Shortly after this time, sunspot AS began to
shrink in size and slowly disintegrate the 5configuration. The last
image in Figure 4.7 is the last observation frame in which any evidence of
spot AS can be seen; it is the slight bulge on the southeast side of
sunspot A.
It would have been interesting to observe the 5spot disintegration
phase of the active regions evolution near the limb in soft Xrays to see
the magnetic loop structure of the flux tube associated with spots AS and
AD. Observations by Lites et al. (1995) of a small 5-spot group in 1992
indicated that they had observed the ascent of a nearly closed magnetic
system through the photosphere into the corona. Four days after the
disappearance of their 5spot, Lites et al. (1995) noted a large Xray loop
structure emanating from the region and extending considerably higher
than the previous emissions which they hypothesized to be the magnetic
loop structure from the 5spot. Based on the white light images
Figure 4.8 A diagram of
the possible sub-photospheric
structure of the flux tubes
which defined spots A and A8.

examined here, this same scenario of a rising closed magnetic loop could
be envisioned to explain the behavior of spots A5 and AD.
Penumbrae Formation
The SOONSPOT white light images were also used to determine
when sunspot penumbrae formed, relative to umbral area, for two of the
dominant spots in AR 7978. The sunspots used in this analysis are spots
B and D. Sunspot A was not chosen because of the complications
associated with defining the umbral area of the spot when it is made up
of two umbrae of opposite polarity. Sunspot C was not used because it
developed its penumbra during the data gap between 8 July and 9 July.
Due to the small number of samples, a statistical analysis can not
be performed to conclude a general relation between umbral area of a
pore and the formation of a penumbra that defines a sunspot. However,
the information presented here can be used in future studies and
observations to, hopefully, provide a better understanding of this
The following sections presented here discuss my methods for
calculating sunspot umbral area and for determining when a penumbra
has formed. In the final section, I present my results and discuss how
they compare to current theories on the transition from pore to sunspot.

Measurements of Penumbral Formation
With the variable seeing of the SOONSPOT white light images, it
is difficult to simply visually examine the images and note when a
penumbra forms for a particular sunspot. To overcome this difficulty, I
used a procedure recommended by Leka (1997) which is less sensitive to
bad seeing than visual inspection. In this process, I plotted intensity
histograms of sub-images which contained only the sunspot being
examined. For an image which only contains a pore, the histogram
shows two distribution peaks one for the quiet sun intensity, and one
for the pore. As a penumbra develops around the pore, a third
distribution peak forms between the first two. Figures 4.9 and 4.10 show
the images and histograms used to determine the time of penumbral
formation for sunspots B and D in this analysis. The histograms are of
the images shown in the figures which are four times the size of the raw
subimages and smoothed to remove noise.
Once the time of penumbral formation was determined for both
spots, two sub-images of sunspots B and D were extracted. The first
subimage is from the frame just before the penumbral formation, and
the second is from the next frame with the penumbra, for each spot.
Next, the number of pixels with umbral intensity, or less, was
determined to define the umbral area of each subimage, in pixels. The
final step was to convert the umbral areas in pixels to areas in square
kilometers. This step required calculating the number of kilometers in
each dimension of a single pixel, based on the pixel resolution, the
diameter of the Sun in arcseconds and in kilometers, and by taking into

0 50 100 150 SOO S50 300
Pixel Value
-S 400
0 50 100 150 SOO S50 300
__________________________________Pixel Value__________________________________
Figure 4.9 Intensity histograms of sub-images of sunspot B just prior to (a) and
after (b) the formation of a penumbra, indicating the penumbral distribution peak
used to determine the time of formation of the penumbra.
i ii|iii i quiet sun
intensity -
initial 1| -
penumbra jW
_ umbra/pore ^ -
. intensity . .i. /TT*iTy i i L ~

50 100 150 POO
Pixel Value
ffi 1500
0 50 100 150 200 250 300
______________________________________Pixel Value_____________________________________
Figure 4.10 Intensity histograms of sub-images of sunspot D just prior to (a) and
after (b) the formation of a penumbra, indicating the penumbral distribution peak
used to determine the time of formation of the penumbra.

quiet sun

account the fact that the heliocentric distance of the x dimension of a
pixel varies as the cosine of the longitude.
Results of Penumbral Formation
The results of the variation of umbral area during penumbra
formation for spots B and D are shown in Table 4.1. The decrease in area
Sunspot B D
Pre-penumbral Area (km2) 1.04 x 108 2.07 x 108
Post-penumbral Area (km2) 1.37 x 10 8 1.60 x 108
Percentage Change + 32% - 23%
Table 4.1 Calculated area of sunspots B and D, just before and after they formed
penumbrae, and the percentage change in area with respect to pre-penumbral
of spot D makes physical sense if the penumbra is formed in response to
an increase in magnetic field strength. If it is assumed that magnetic flux
is conserved through the photosphere-chromosphere boundary during
the 30 minute time period between white fight images, then the
penumbra formation of sunspot D could result from fallen flux tubes as
described by Wentzel (1992). In this paper, Wentzel (1992) suggests that
an increase in the vertical magnetic field strength of a pore causes
individual flux tubes to fall away from the center and lie horizontally
near the surface, creating the penumbra (Figure 4.11). However,
Rucklidge, Schmidt, & Weiss (1995) present a theoretical argument,

Inclined Field Lines
Inclined Field Lines
Figure 4.11 A diagram illustrating two competing penumbral formation models.
The top diagram illustrates the theory fallen flux tubes due to increased magnetic field
strength by Wentzel (1992). The bottom diagram shows penumbral formation from
ascending field lines which are already inclined, as theorized by Rucklidge, Schmidt,
& Weiss (1995).
based on the transport of energy across the magnetopause boundary of
the umbra, which states that the inclination from the vertical of the
emerging magnetic field of a pore is the key factor for determining the
formation of a penumbra. Their conclusion is that the horizontal fields
rise through the photosphere already inclined from the vertical and forms
the penumbra from below the solar surface, rather than from field lines
above the surface. This conclusion is also supported by Stokes vector
magnetogram observations by Leka (1997) in which no gradual increase

in the inclination angle was observed in a pore to sunspot transition,
suggesting that the highly inclined penumbral field emerged from below
the photosphere. Therefore, it is possible that sunspot D formed its
penumbra in response to an increase of the local magnetic field strength
which emerged in a highly inclined orientation.
The theory of the inclined penumbral field emerging through the
photosphere would also explain the formation of sunspot Bs penumbra, if
conservation of flux is assumed. The formation of the penumbra of
sunspot B can not be explained by an increase in magnetic field strength
because of its increase in umbral area, since this would violate the flux
conservation law. However, if the penumbral formation is dependent on
the inclination of the field lines, rather than the field strength, then the
increase in umbral area is not of consequence. However, if there is a
increase in field strength of spot B from the coalescence of pores, the
excess flux from the smaller umbral area must be accounted for in the
newly formed penumbra, assuming flux is conserved.

This chapter presents my analysis of SOONSPOT H-a and Yohkoh
soft X-ray observations. The goal of this investigation is to test the
nonlinear numerical simulation models of Shibata et al. (1989), Shibata,
Nozawa, & Matsumoto (1992), and Matsumoto et al. (1993) which
describe the ascent of an active region magnetic flux tube from beneath
the photosphere to the lower corona. The claim of these simulations is
that flux tube emergence through the chromosphere should show the
disappearance of a filament in H-a and a spatial and temporal
correspondence of Xray bright points. To test these numerical models
requires co-aligning H-a and soft X-ray images in space and time during
an identifiable event of a disappearing filament.
In the first section of this chapter, I present an overview of the
numerical simulation models of Shibata et al. (1989), Shibata, Nozawa, &
Matsumoto (1992), and Matsumoto et al. (1993) and discuss their
numerical integration technique. The following section describes the
filament disappearance events that were chosen and the limitations of
the data. In the last sections of this chapter, I discuss the technique used
to coalign the Yohkoh and SOONSPOT images, and present my results

of the testing of the numerical simulations.
Numerical Simulations and Models
The nonlinear numerical simulations of emerging magnetic flux in
the solar atmosphere that are being tested in this analysis are those of
Shibata et al. (1989), Shibata, Nozawa, & Matsumoto (1992), and
Matsumoto et al. (1993). The first two references are simulations in two-
dimensional (2D) space, one horizontal dimension (x) and the dimension
vertical to the solar surface (z), while that of Matsumoto et al. (1993) is a
three-dimensional (3D) simulation. Each of these models begins with
an initial flux tube, localized in the convection zone at the bottom of the
photosphere, which is unstable to the undular mode of the magnetic
buoyancy instability (Shibata, Nozawa, & Matsumoto, 1992). Small
velocity perturbations are initially imposed on the flux tube, and the
system is integrated in space and time to a point where magnetic
reconnection occurs with the coronal magnetic field.
Numerical Methods and Predictions
The model solar atmosphere used in all three of these simulations
covers a vertical range from the lower photosphere, through the
chromosphere, up to the lower corona. The horizontal dimension(s) is
(are) scaled according to the initial conditions placed on the initial
magnetic flux tube and the mode of the instability. The models assume

that the medium is an ideal gas and that the magnetic field is frozen in
the gas (Matsumoto et al., 1993).
The physical properties of emerging flux anticipated by these
numerical models are many. However, the predicted features examined
by the data in this study focus specifically on the claim that bright H-a
plage and X-ray bright points should coincide with the disappearance of
a filament by gravitational draining of the flux tube, and possibly a
filament eruption (Shibata et al., 1989; Shibata, Nozawa, & Matsumoto,
1992; Matsumoto et al., 1993). Figure 5.1 is a simplified diagram which
summarizes this process predicted to cause these events.
In diagram (a) of Figure 5.1, the flux tube is shown as it initially
penetrates through the photosphere into the chromosphere. As it
emerges, pores, or sunspots, are formed in the photosphere at the
footpoints of the flux tube. The plasma contained in the rising flux tube
is cooler than the surrounding chromospheric gas and appears dark
compared to the local environment, forming an arch filament. As the flux
tube continues to rise due to magnetic buoyancy (Figure 5.1 b), it expands
horizontally to account for the decreased pressure in the chromosphere.
At this point, the material within the flux tube begins to gravitationally
drain down the field lines at three to five times the local sound speed (SO-
SO km/s), creating strong MHD shock waves at the footpoints (Shibata et
al., 1989; Matsumoto et al., 1993). In the final stage, depicted in Figure
5.1 c, the flux tube rises to the base of the corona and a current sheet is
created between the flux tube field and the coronal field, while energetic,
fast magnetic reconnection begins to take place (Shibata, Nozawa, &

Rising Flux Tube
^__________ Sunspots/Pores
Photosphere/Chromosphere Interface

Photosphere/Chromosphere Interface
Coronal Magnetic Field
Photosphere/Chromosphere Interface (c)
Figure 5.1 Simplified sketches illustrating the rise of a magnetic flux
tube as predicted by the nonlinear numerical simulations of Shibata et
al. (1989), Shibata, Nozawa, & Matsumoto (1992), and Matsumoto et
al. (1993).

Matsumoto, 1992). Within the current sheet, magnetic islands are
formed, due to multiple X line reconnection, which confine cool, dense,
high pressure chromospheric plasma (Shibata et al., 1989; Shibata,
Nozawa, & Matsumoto, 1992). Both the high temperature current sheet
plasma and the low temperature magnetic island plasma are accelerated
along the current sheet at speeds up to the Alfven speed (50-100 km/s),
generating fast shocks at the edges of the current sheet (Shibata et al.,
1989; Shibata, Nozawa, & Matsumoto, 1992). Depending on the specific
geometry of the flux tube field and the coronal field (parallel, antiparallel,
or something in between, or a vertical component to the coronal field), the
flow of the cooler, and therefore darker, magnetic island plasma would be
observed as either an outward surge, or a downward flow similar to the
gravitational draining of the flux tube (Shibata et al., 1989; Shibata,
Nozawa, & Matsumoto, 1992).
At the footpoints of the rising flux tube and the interface between
the flux tube and the corona, the locations where MHD shocks are
formed, all three of these simulations predict these locations to be sources
of Xray bright points. Therefore, observationally, Xray bright points
should coincide, in space and in time, with the disappearance of an H-a
filament in an emerging flux region.
Testing of Nonlinear Numerical Simulations
To test the numerical simulations described above required the co-
alignment of Yohkoh soft Xray images of AR 7978 with the SOONSPOT

H-a images, for disappearing filament events observed in H-a. The first
step in this process involved identifying events of disappearing filaments
from the SOONSPOT images and then correlating the observation times
with the soft X-ray data. After events with simultaneous coverage were
determined, images from the two data sets were spatially co-aligned,
based on matching of pixel coordinates, to test the predictions of the
numerical simulations. The sections that follow discuss the techniques
used in this process, and the results of the investigation.
Selection of Events
Due to AR. 7978s rapid formation and its 6-spot configuration, the
H-a images were dominated by bright plage and flare kernels which did
not make it possible to reliably determine the presence and gas motions
of filaments which were less than about 20 pixels in length. Therefore,
only large scale active region filaments were examined from this data set.
The process for determining an event of a disappearing filament
was a visual inspection of computer generated movies of chronologically
ordered SOONSPOT H-a image sequences. To account for slight
differences in telescope pointing of the various observatories, the image
sequences were first coaligned using a crosscorrelation technique on a
small region of the images which contained somewhat long lasting
features. The effect of this process was to remove image by image
jumping of the active region features to produce a relatively jitter free

From the three days worth of H-a images used in this study, eight
obvious events of disappearing filaments were noted. Of these eight
events, three of them were correlated in time with the Yohkoh soft X-ray
images and were chosen for this investigation. These three events all
occurred on 7 July from 13:00 to 13:57 UT, 14:04 to 14:20 UT, and 14:43
to 14:54 UT. Two other events simultaneously occurred on 9 July,
however, most the Yohkoh soft X-ray images for the time period were
saturated over the entire active region and could not be used, excluding
this data set. Data from 8 July was excluded due to frequent flaring
which saturated the active region in both the H-a and soft X-rays
Coalignment of Images
The process for coaligning Yohkoh soft Xray images and
SOONSPOT H-a images required only a few steps of preparation and the
use of a modified version of the COAL_IMAGE IDL procedure from the
Yohkoh software library. The first step in this procedure was to extract
partial frame images of AR 7978 from the Yohkoh full disk X-ray images
for the time period of each event noted above. This was easily
accomplished with the use of the Yohkoh library procedure
EXT_SUBSET which uses user selected image coordinates to extract 64 x
64 pixel subimages from the full disk images. The output data
structures in IDL contains all of the relevant information for processing
the images, such as telescope pointing, date and time of observations,
heliocentric coordinates of each image, exposure time, and pixel

resolution. The final step in preparing the X-ray images for co-
alignment involved the use of the SXT_PREP IDL procedure which aligns
all of the images in a partial frame data set, normalizes them, subtracts
the dark current, and rotates the data so that solar north is at the top of
each image.
For the SOONSPOT H-a images, the only preparation that was
needed was the determination of the distance, in pixels, from the solar
disk center to the lower left corner pixel of each image, and a slight image
rotation to remove the solar P angle. The pixel coordinate values were
calculated by the method described in chapter 5 from the reference pixel
defined in the SOONSPOT image headers, and the rotation of the images
was performed by the IDL ROT procedure.
Once the images from both data sets were prepared, the co-
alignment of nearly simultaneous images was performed with a modified
version of the IDL procedure COAL_IMAGE from the Yohkoh software
library. The COAL_IMAGE procedure enlarges the lower resolution
Yohkoh images (9.8 arcsec/pixel) to the same scale as the SOONSPOT
images, based on pixel resolution, and then extracts a subimage which
has its lower left corner pixel aligned with the lower left corner pixel of
the SOONSPOT images. The resulting image that is returned by
COAL_IMAGE is one that is the same scale and size (512 x 512) of the
SOONSPOT images. The modification that was made to the
COAL_IMAGE procedure was the addition of an extra input variable for
performing scaling of the Yohkoh images in the ydirection separately
from the xdirection. The original COAL_IMAGE procedure assumed
that the ground based images (SOONSPOT) would have square pixels

and scaled the Yohkoh images equally in both the dimensions. This
modification allowed the procedure to account for the rectangular shape
of the SOONSPOT pixels, and to accurately perform the coalignment of
the images.
Test Results
The results of comparing the coaligned soft X-ray and H-a
images, for testing of the nonlinear numerical models mentioned above,
are not discouraging. However, based on the data presented in this
study, it is evident that the numerical simulations of Shibata et al.
(1989), Shibata, Nozawa, & Matsumoto (1992), and Matsumoto et al.
(1993) are not correct in all aspects. In the following discussion, I
illustrate, by way of condensed time series images of the soft Xray and
H-a data, that the emergence of new magnetic flux through the
photosphere does not necessarily require the disappearance of an H-a
filament and the production of strong X-ray emissions. In fact, the
results of the three events examined in this investigation suggest that
new emerging magnetic flux is associated with either strong Xray
emission or the disappearance of an H-a filament, but not both.
For reference, Figure 5.2 contains the available white light images
of AR 7978 for each of the time sequences. There are two regions of
interest to note in these images. The first is the area around sunspot C
(same labeling as in chapter 5) in which flux is coalescing to form spot C,
and the second is the small grouping of pores near the bottom of the
images, between sunspots A and D.

The first filament
disappearance event that
was analyzed occurred
between 13:00 and 13:57
UT on 7 July. The event is
illustrated in Figure 5.3
which contains selected H-
a images from this time
sequence. The white boxes
in several of the frames
indicate the images that
were coaligned to
corresponding soft Xray
observations which are displayed in Figure 5.4. The white box itself
outlines the region of the disappearing filament which is also the area of
the emerging pores to the south and east of sunspot D. The black box in
the first frame of Figure 5.3 indicates the region where the Xray images
show the strongest emission which is centered in the area around sunspot
C. In essence, these sequences of X-ray and H-a images are showing
strong X-ray emission and the disappearance of a filament over regions
of emerging flux, but not the same regions. It is also worth noting that
the arch filaments that are present in the area around spot C and the
enhanced Xray emission show no distinguishable change in size or shape
throughout the observations.
The second event, illustrated in Figures 5.5 and 5.6, shows similar
results as the first. The strong Xray emission is centered over the
m m 1 %

7/07/98 13:01.-32 7/07/98 13:33:32
%. %
7/07/96 14:03:32 7/07/go 14:33:32
* SoJar North
7/07/96 15:03:32
Figure 5.2 SOONSPOT white light images
corresponding to the time sequences of the three
disappearing filament events observed in H-a.

/90 13:00:19 7/07/90 13:0?:48 7/07/98 13:05:19 7/07/98 13:10:19 7/07/96 13:15:1!
790 13:15:19 7/07/90 13:3?: 19 7/07/98 13:37:19 7/07/98 13:47:19 7/07/98 13:57:1!
.3 An abbreviated time series of AR 7978 which displays the disappearance of an H-a filament over a reg
dng magnetic flux. The images with white boxes outlining the region of interest were coaligned with nearly
leous soft Xray images. The black box in the first frame indicates the location of strongest Xray emission. r
1 date of each observation is given below each image.

Figure 5.4 The Yohkoh soft X-ray images associated with the H-a images in Figure 5.3. The
white boxes indicate the region of interest where the event of a disappearing filament occured.
The time and date of each observation, along with the exposure time, is given below the individual

7/07/98 14:03:49 7/07/98 14:08:49
7/07/93 14:12:49 7/07/93 14:17:49
7/07/98 14:19:49
Figure 5.5 A condensed time series of Ha images, illustrating
the second event of a disappearing filament examined in this
investigation. The white boxes outline the region of interest, and
the black box in the first frame represents the location where the
brightest Xray emissions were observed.

7-JUL-98 14:02:4] 688.0 ms 7-JUL-96 14:06:57 38.0 ms
7-JUL-98 14:11:13 698.0 ms 7-J1JL-96 14:15:29 38.0 ms
7-JUL-98 14:19:45 688.0 ms
Figure 5.6 Yohkoh soft X-ray images corresponding to the
H-a images in Figure 5.5. The white box indicates region where
the second disappearing filament event occured.

region around sunspot C, and the disappearing filament occurs between
sunspot D and the grouping of pores to the south. As with the first event,
the H-a filaments that are associated with the region of enhanced X-ray
emission and sunspot C show no change throughout the time sequence.
In fact, these filaments do not show any discernible change from the
previous event one hour earlier.
The most obvious event occurred between 14:43 and 14:54 UT in
which a filament, again in the vicinity of sunspot D, erupted. Figure 5.7
shows an abbreviated time series of this event in H-a and Figure 5.8
shows the corresponding soft Xray images. As with the previous
images, the white boxes outline the region of interest where the filament
eruption occurred, and the black box in the first H-a image indicates the
location of the strongest Xray emission.
Once again, the majority of the enhanced X-ray emission comes
from the region around sunspot C with still no observable change in the
filaments of this area. There is, however, an X-ray bright point just
outside the northeast corner of the white box surrounding the filament
region of interest which is well defined in the short exposure (38 ms)
Yohkoh image at 14:49 UT. This X-ray bright point is most likely
associated with the bright flare kernel, seen in H-a in the same region,
which may also be the cause of the filament eruption.
From the analysis of the observations of these three events, it is
evident that emerging flux is associated with either strong Xray
emission or the disappearance of an H-a filament, but not necessarily
both. The data from the third event indicates that a third element, an H
a flare, may be necessary for these two signatures of emerging flux to

7/07/90 14:43:19 7/07/90 14:45:19
' " ' : V '$* ^ ,
7/07/90 14:47:19 7/07/90 14:49:49
7/07/90 14:51:49 7/07/90 14:53:49
Figure 5.7 An erupting filament event in H-a. The white box
outlines the event in the images, and the black box indicates the
location where the brightest Xray emission was observed. Images
1, 2, 4, and 6 were coaligned with the soft Xray images in Figure
5.8. Note the bright flare kernel near the northeast corner of the
white box which corresponds to an X-ray bright point in Figure 5.8.

Figure 5.8 Yohkoh soft X-ray images which were coaligned
with the SOONSPOT H-a images in Figure 5.7. The white boxes
correspond to the location of an erupting filament in H-a.

occur simultaneously in space and in time. If this is the case, then the
numerical simulations of Shibata et al. (1989), Shibata, Nozawa, &
Matsumoto (1992), and Matsumoto et al. (1993) would be better suited to
describe a method for the occurrence of solar flares.

My analysis of the formation of NOAA AR 7978 was, to say the
least, interesting. Its quick emergence from quiet Sun to a
8configuration made it an excellent active region to study in terms of its
dynamics of growth and provided the opportunity to test the numerical
simulations of Shibata et al. (1989), Shibata, Nozawa, & Matsumoto
(1992), and Matsumoto et al. (1993).
This chapter presents conclusions on the growth and formation of
AR 7978. The first section discusses results of sunspot proper motions
and penumbrae formation, with respect to the white light observations.
The next section presents the results of testing the nonlinear numerical
simulations of Shibata et al. (1989), Shibata, Nozawa, & Matsumoto
(1992), and Matsumoto et al. (1993) and suggests a possible explanation
for the differences between the numerical predictions and the H-a and
soft X-ray observations. In the last section, I present a closing statement
with regard to the data, my analysis of the data, and future studies on
the topics presented in this thesis.

White Light Observations
The SOONSPOT white light images provided adequate coverage of
AR 7978 for determining sunspot proper motions of the region. However,
the variable seeing, the large field of view of the large scale images, and
the 12 hour data gaps between observing days did lead to difficulties and
uncertainties some of the calculations, especially for determining when
penumbrae formed around spots. The following sections discuss my
conclusions about the dominant sunspot proper motions, and those
related to the spots of the 5-configuration. The findings for penumbral
formation for two of the dominant sunspots are also reiterated here.
Sunsnot Proper Motions
The relative motions of the dominant sunspots of AR 7978 showed
no extraordinary variations from what has been observed of sunspots in
the past. Based on the analysis performed here, the leading p-spot, D,
showed an overall two and a half day change in longitudinal position of
0.6 which is gives average rate of 0.25 per day. The maximum rate of
sunspot D was 0.8 per day and occurred in the first 24 hour period of
observations. The trailing f-spots, B and C, showed an overall
longitudinal change in position of 1.4 and 1.2, respectively. Sunspot B
had its maximum rate of motion on the first day of observations, with an
average longitudinal rate of 1 per day. Sunspot C showed its maximum
rate of motion on 8 July with an average rate of 1 per day.
From these measurements, it is apparent that the majority of

sunspot motion occurs as the flux tube initially ascends into the
chromosphere, and begins to interact with the coronal magnetic field.
This suggests the possibility that the initial sunspot motions may be due
to forces on the emerged section of the flux tube from the chromosphere
and corona (i.e., plasma motions, ambient magnetic fields, and pressure
gradients) which act in such a way as to achieve an equilibrium between
the flux tube and the new environment. In this scenario, the sunspots
would then represent the boundary between the subphotospheric
structure of the flux tube, influenced by convective forces from below, and
overlying atmospheric forces from the chromosphere and corona. It
would be interesting to continue this study and investigate the balance of
forces, in theoretical detail, from the photospherechromosphere
boundary to the corona with a nonstatic atmosphere.
Delta-Spot Proper Motions
The proper motions of the sunspots related to the 5-spot displayed
an interesting sequence of events. First, the nearly simultaneous
rotation of spots A5 and AD about sunspot A, and the observable
intermixing of their penumbrae, indicates there was a direct, rigid
magnetic link between these two sunspots. The rotation itself, suggests a
twist between the flux tubes of the two opposite polarity umbrae of the
5spot, ascending from below the photosphere. Second, in the final
stages of the 5-configuration, the penumbrae connecting spots A5 and AD
are observed pulling apart as the two spots begin to separate. Then,

within two hours of their observed separation, an X2.6 flare was observed
from the region of the neutral line that separated spots A8 and AD, and
the 5-configuration quickly disintegrated shortly after this event. This
immense flare, the largest in several years, indicates a major re-
organization of the magnetic field in this area of the active region, and
suggests a reconnection event between spot AD and the overlying field,
leaving spot A5 unconnected to any non-5-configuration spots which
resulted in its destruction. The white light and Ha history of this event
are displayed in Figure 6.1.
Penumbral Formation
My investigation of penubral formation versus umbral area is
inconclusive due to only having a sample of two sunspots to study. In the
two pore to sunspot transitions that were studied in this thesis, it was
found that the umbral area of one spot increased with the formation of a
penumbra, while the umbral area of the other decreased. This implies
that penumbral formation is not necessarily due to an increase in
magnetic field strength through the surface of a pore, and that the
theories of highly inclined fields emerging through the photosphere to
form penumbrae may be more correct. The latter interpretation would
explain the two events observed in this investigation.

9/07/96 08:53:33
9/07/96 09:24:33
9/07/96 09:40:29
Figure 6.1 The evolution of the X2.6 flare from AR 7978 in H-a. The top two
images are the corresponding white light images during the flare.

Hydrogen Alpha and Soft X-Rav Observations
The H-a and soft X-ray images of AR 7978 were also used to test
the prediction of the nonlinear numerical simulations of Shibata et al.
(1989), Shibata, Nozawa, & Matsumoto (1992), and Matsumoto et al.
(1993) that new emerging flux is accompanied by the disappearance of an
H-a filament and X-ray bright points. Shibata et al. (1989), Shibata,
Nozawa, & Matsumoto (1992), and Matsumoto et al. (1993) use their
calculated velocity values to determine locations of MHD shocks which
they claim to be sources of Xray bright points and associated with the
gravitational draining/disappearance of an H-a filament in an emerging
flux region. However, based on the observations presented here, these
two processes do not, in general, occur co-spatially. The events
investigated in this thesis show that disappearing Ha filaments and
X-ray bright points are both associated with emerging flux, but not
necessarily in the same spatial location, or with the same emerging flux
region. Of the three events studied here, only the third one (7 July
14:43-14:54 UT) showed any indication of spatial alignment of these two
signatures and was accompanied by a localized flare. This evidence
indicates that the physical assumptions of the models are insufficient to
provide a solid basis for accurately determining plasma motions, and
therefore X-ray bright points and filament disappearance, in an
emerging flux region. Therefore, it is concluded that these numerical
models do not, in general, properly describe magnetic flux tube
emergence in an active region, and its ascent through the solar

Closing Statement
The data used in this investigation of the growth and formation of
NOAA AR 7978 were adequate for the study presented here. However, to
achieve more accurate results, higher temporal and spatial resolution
would be beneficial. With the full operation of all five SOONSPOT
observatories, the success of the Solar and Heliospheric Observatory
(SOHO), and the capabilities of Stokes vector magnetograms, it should be
possible to obtain simultaneous 24 hour coverage of an emerging active
region in a wide range of wavelengths for a complete and total analysis of
emerging flux through the solar atmosphere. As coordinated
observations between these instruments, and others, occur, I believe that
more accurate theoretical models will develop and that we will have a
much better understanding of the solar atmosphere, the Suns magnetic
field, and the effects solar events have on the Earths environment.

Ambastha, Ashok, Mona J. Hagyard, & E. A. West, Evolutionary and
Flare-Associated Magnetic Shear Variations Observed in a
Complex, Flare-Productive Active Region, 1993, Solar Phys., 148,
pp. 277-299.
Bransden, B. H. and C. J. Joachain, Physics of Atoms and Molecules.
Longman Scientific & Technical, Essex, England, 1983.
Canfield, Richard C., Alan L. Kiplinger, Matthew J. Penn, & JeanPierre
Wiilser, Ha Spectra of Dynamic Chromospheric Processes in Five
Well-Observed X-Ray Flares, 1990, Ap. J., 363, pp. 318-325.
Chen, Francis F., Introduction To Plasma Physics and Controlled Fusion.
Plenum Press, New York, 1984.
Chen, Jimin, Haimin Wang, Harold Zirin, & Guoxiang Ai, Observations
of Vector Magnetic Fields in Flaring Active Regions, 1994, Solar
Phys., 154, pp. 261-273.
De La Beaujardiere, J. F. Richard C. Canfield, & K. D. Leka, The
Morphology of Flare Phenomena, Magnetic Fields, and Electric
Currents in Active Regions. III. NOAA Active Region 6233 (1990
August), 1993, Ap. J., 411, pp. 378382.
De La Beaujardiere, J. F., Alan L. Kiplinger, & Richard C. Canfield,
Coordinated Spectral and Temporal Ha Observations of a Solar
Flare, 1992, Ap. J., 401, pp. 761-767.
Demoulin, P., L, van DrielGesztelyi, B. Schmieder, J. C. Henoux, G.
Csepura, & M. J. Hagyard, Evidence for Magnetic Reconnection in
Solar Flares, 1993, Astron. Astrophys., 271, pp. 292-307.

Droge, Wolfgang, Peter Meyer, Paul Evenson, & Dan Moses, Electron
Acceleration in Solar Flares, 1989, Solar Phys., 121, pp. 95103.
Gaizauskas, V., K. L. Harvey, and M. Proulx, Interactions Between
Nested Sunspots. I. The Formation and Breakup of a Delta-Type
Sunspot, 1994, Ap. J., 422, pp. 883-898.
Gary, D. E., Y. Leblanc, G. A. Dulk, & L. Golub, The Magnetic Field in
the Corona Above Sunspots at the Eclipse of 1991 July 11, 1993,
Ap. J 412, pp. 421-430.
Gibson, Edward G., The Quiet Sun. U. S. Government Printing Office,
Washington, D. C., 1973.
Gosling, J. T., The Solar Flare Myth, 1993, J. Geophys. Res., 98, pp.
Hartman, William K., Astronomy: The Cosmic Journey. Wadsworth
Publishing Company, Belmont, CA, 1991.
Hecht, Eugene, Optics, second edition. Addison-Wesley Publishing
Company, Reading, MA, 1987.
Hongqi, Zhang, Structures of Chromospheric Magnetic Fields in the
Solar Flare Producing Active Region 5747, 1993, Solar Phys., 144,
pp. 323-340.
Howard, Robert F., The Growth and Decay of Sunspot Groups, 1992,
Solar Phys., 137,pp. 5165.
Jones, Harrison P., Thomas L. Duval, JR., John W. Harvey, Charles T.
Mahaffey, Jan D. Schwitters, & Jorge E. Simmons, The
NASA/NSO Spectromagnetograph, 1992, Solar Phys., 139, 211-
Judge, Philip, 1994, An Introduction To The Solar Atmosphere,
unpublished lecture notes.
Leka, K. D., 1997, personal communication, Discussion of penumbra

Leka, K. D., Richard C. Canfield, A. N. McClymont, J. F. De La
Beaujardiere, Yuhong Fan, & F. Tang, The Morphology of Flare
Phenomena, Magnetic Fields, and Electric Currents in Active
Regions. II. NOAA Active Region 5747 (1989 October), 1993, Ap.
J., 411, pp. 370-377.
Lites, B. W., D. F. Elmore, P. Seagraves, & A. P. Skumanich, Stokes
Profile Analysis and Vector Magnetic Fields. VI. Fine Scale
Structure of a Sunspot, 1993, Ap. J., 418, pp. 928942.
Lites, B. W., B. C. Low, V. Martinez Pillet, P. Seagraves, A. Skumanich,
Z. A. Frank, R. A. Shine, & S. Tsuneta, The Possible Ascent of a
Closed Magnetic System Through the Photospher, 1995, Ap. J.,
446, 877-894.
Matsumoto, R., T. Tajima, K. Shibata, & Kaisig, Three-Dimensional
Magnetohydrodynamics of the Emerging Magnetic Flux in the
Solar Atmosphere, 1993, Ap. J., 414, 357-371.
Mazzucconi, F., C. Coveri, & G. Godoli, Sunspot Proper Motions and
Behaviour of Rising Magnetic Flux Tubes, 1990, Solar Phys., 125,
pp. 269-276.
Morrison, M., Yohkoh Analysis Guide. Volume 1: Users Guide. Lockheed
Palo Alto Research Laboratory, 1994.
Morrison, M., Yohkoh Analysis Guide. Volume 2: Reference Guide,
Lockheed Palo Alto Research Laboratory, 1994.
Morrison, M., Yohkoh Analysis Guide, Volume 3: Instrument Guide,
Lockheed Palo Alto Research Laboratory, 1994.
Parks, George K., Physics of Space Plasmas, An Introduction, Addison-
Wesley Publishing Company, Redwood City, CA, 1991.
Rubin, Ephraim L., Samuel Z. Burstein, Difference Methods for the
Inviscid and Viscous Equations of a Compressible Gas, 1967, J.
Comp. Phys., 2, 178196.

Rucklidge, A. M., H. U. Schmidt, & N. O. Weiss, The Abrupt
Development of Penumbrae in Sunspots, 1995, Mon. Not. R.
Astron. Soc., 273, 491-498.
Shibata, K., T. Tajima, R. S. Steinolfson, & R. Matsumoto, Two-
Dimensional Magnetohydrodynamic Model of Emerging Magnetic
Flux in the Solar Atmosphere, 1989, Ap. J., 345, 584596.
Shibata, Kazunari, Satoshi Nozawa, & Ryoji Matsumoto, Magnetic
Reconnection Associated with Emerging Magnetic Flux, 1992,
Publ. Astron. Soc. Japan, 44, 265-272.
Skumanich, A., B. W. Lites, B. C. Low, & V. Martinez Pillet, to appear in
La polarimetrie, outil pour letude de lactivite magnetique solaire
et stellaire, 1995, N. Mein & S. Sahal (eds.) Obs. de Paris, Meudon.
Smith, Jr., Jesse B., Donald F. Neidig, Philip H. Wiborg, Edward A.
West, Mona J. Hagyard, Mitzi Adams, & Paul H. Seagraves, An
Objective Test of Magnetic Shear as a Flare Predictor, 1996, Solar
Drivers of Interplanetary and Terrestrial Disturbances, ASP
Conference Series, Vol. 95, K. S. Balasubramaniam, Stephen L.
Keil, & Raymond N. Smartt (eds.), pp. 55-65.
Snow, Theodore P., Essentials of The Dynamic Universe. An Introduction
to Astronomy, second edition, West Publishing Company, St. Paul,
MN, 1987.
Solanki, Sami K., Small-Scale Solar Magnetic Fields: An Overview,
1993, Solar Phys., 63, pp. 1-188.
Solanki, S. K., C. A. P. Montavon, & W. Livingston, Infrared Lines as
Probes of Solar Magnetic Features VII. On the Nature of the
Evershed Effect in Sunspots, 1994, Astron. Astrophys., 283, pp.
Tanaka, Katsuo, Studies on a Very Flare-Active 5 Group: Peculiar 8 Spot
Evolution and Inferred Subsurface Magnetic Rope Structure,
1991, Solar Phys., 136, pp. 133-149.

The Astronomical Almanac for the Year 1996, Particle Physics and
Astronomy Research Council, U. S. Government Printing Office,
Washington, D. C., 1995.
Tsuneta, S., L. Acton, M. Bruner, J. Lemen, W. Brown, R. Carvalho, R.
Catura, S. Freeland, B. Jurcevich, M. Morrison, Y. Ogawara, T.
Hirayama, & J. Owens, The Soft X-Ray Telescope for the SOLAR-
A Mission, 1991, Solar Phys., 136, 3767.
Wentzel, Donat G., Structure of Sunspot Penumbrae: Fallen Magnetic
Flux Tubes, 1992, Ap. J., 388, pp. 211217.
Winglee, R. M., A. L. Kiplinger, D. M. Zarro, G. A. Dulk, & J. R. Lemen,
Interrelation of Soft and Hard Xray Emissions During Solar
Flares. I. Observations, 1991, Ap. J., 375, pp. 366381.
Zhang, Guiquing, An Evolutionary Feature of the Magnetic Field Related
to Flares in Active Regions, 1994, Solar Phys., 154, pp. 41-49.
Zirin, Harold, Astrophysics of the Sun. Cambrige University Press, New
York, 1988.
Zirin, Harold and Frances Tang, Optical Properties of Impulsive Flares,
1990, Ap. J., 73, pp. 111-115.
Zirin, Harold and Haimin Wang, Strong Transverse Fields in d-Spots,
1993, Solar Phys., 144, pp. 37-43.