Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://star.arm.ac.uk/preprints/441.ps Äàòà èçìåíåíèÿ: Wed May 11 18:53:13 2005 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 21:28:31 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: mdi

Astronomy & Astrophysics manuscript no. 3099 May 11, 2005
(DOI: will be inserted by hand later)
Evidence for Explosive Event Activity Originating in the
Chromosphere
J.G. Doyle 1 , B. Ishak 1 , I. Ugarte­Urra 1# , P. Bryans 2 , and H.P. Summers 2
1 Armagh Observatory, College Hill, Armagh BT61 9DG, N. Ireland
2 Department of Physics, University of Strathclyde, Glasgow, Scotland
the date of receipt and acceptance should be inserted later
Abstract. We report on a joint SUMER, CDS, TRACE study, concentrating on a region which shows prolonged
EUV explosive event (EE) activity in the transition region line N v 1238 š A, yet little evidence of such activity in
another transition region line O v 629 š A (formed at a similar temperature) which was observed simultaneously.
A possible explanation for the lack of major activity in the O v line in several explosive events could be that
they originate in the lower chromosphere. This is consistent with the enhancements in the C i 1249 š A line and
with the findings of another study which reported time delays between the chromospheric and transition region
lines in some EE's using high cadence observations (10 s exposure time) obtained with the SUMER spectrometer
in H i Ly 6 (20 000 K) and S vi (200 000 K). Using the generalized collisional­radiative picture, including the
population of metastable levels, we derive the density dependent contribution function for both N v 1238 and
O v 629 for four values of the electron density; 10 6 cm -3 representing the low density limit, 10 9 cm -3 for a typical
quiet Sun electron density plus 10 11 cm -3 and 10 12 cm -3 for an active region. These calculations show that with
increasing electron density, both lines shift to slightly lower temperatures. However, the major di#erence is in the
relative increase in the line flux with increasing density. For N v, increasing the density to 10 11 cm -3 results in
a 60% increase in the line flux, while O v shows a 30% decrease. Increasing the electron density to 10 12 cm -3
results in a factor of two decrease in the O v flux, thus making it di#cult to detect explosive event activity in
this line if the event is formed in the chromosphere. Other explosive events which show simultaneous activity in
both lines are probably formed in the transition region. In one such event, activity is observed in both N v and
O v, yet nothing in C i. In this event we also observe an increase in the TRACE 173 emission, delayed by # 40 s
compared to the transition region lines.
Key words. Sun: corona .. Sun: transition region .. Sun: chromosphere .. Line: formation .. Atomic processes
1. Introduction
High resolution spectral and spatial coverage of the Sun
over the last decade from a range of telescopes aboard
many spacecraft has led to a wealth of observations of
small­scale dynamic events observed from the chromo­
sphere to the transition region and corona. One of the
most observed but yet not understood are explosive events
(EE's), also termed bi­directional jets. They were first dis­
covered and classified as turbulent events by Brueckner
& Bartoe (1983). They are characterized by highly non­
Gaussian profiles due to an enhancement in the blue and
red wings. Most of the events are predominantly blue­
shifted, showing Doppler shifts up to 150 km s -1 . They
appear preferably in regions with weak fluxes of mixed po­
larity or on the border of regions with large concentration
Send o#print requests to: http://star.arm.ac.uk/preprints/
# now at: Naval Research Laboratory, Code 7673U, 4555
Overlook Ave SW, Washington DC 20375, U.S.A
of magnetic flux (Chae et al. 1998). Their average life­
time ranges from # 60 to 350 s. In an analysis of a large
raster, Teriaca et al. (2004) estimated an average size of
1 800 km and a birthrate of 2 500 s -1 over the entire Sun.
They are often observed in bursts lasting up to 30 minutes
in regions undergoing magnetic cancellation (Dere 1994).
Time delays between the chromospheric and transi­
tion region lines have being investigated (Madjarska &
Doyle 2002) using high cadence observations (10 s ex­
posure time) obtained with the SUMER spectrometer in
H i Ly 6 (20 000 K) and S vi (200 000 K). A time delay in
the response of the S vi line with respect to the H i Ly 6
line was observed. This suggests that the process which
generates these jets first occurs in the low chromosphere
although they are best observed in transition region lines.
Teriaca et al. (2001) obtained high spectral resolu­
tion data for a large event in the transition region line
N v 1238 š A and the coronal line Mg x 625 š A. They con­
cluded that the events observed in N v showed a small

2 Doyle et al.: Evidence for UV Explosive Event Activity Originating in the Chromosphere
Fig. 1. Time series image plots for N v 1238 (top) O v 629 (bottom) over the whole duration of the dataset showing the variation
in the line width assuming a single Gaussian fit. The data set was taken on 1999 June 1 starting at 09:13 UT. Below each image
we show the resulting line profile for both O v and N v taken at the same time at five di#erent temporal locations (09:47,
09:53, 10:00, 10:03 and 10:05 UT) labeled (a), (b), (c), (d) and (e), normalized to unity. The quiet Sun profile is over­plotted
for comparison (dashed line).
enhancement in Mg x which was however due to the pres­
ence of a close­by blend from a Si ii line. The conclusion
was that EUV bi­directional jets are not directly relevant
in heating the corona and are characteristic of structures
not obviously connected with the upper corona.
In some instances, a non­Maxwellian electron distri­
bution may be present (Doyle et al. 2003). With a non­

Doyle et al.: Evidence for UV Explosive Event Activity Originating in the Chromosphere 3
Maxwellian electron distribution, a significant fraction of
the emission typical for a plasma at a temperature just be­
low 1 MK can instead originate at temperatures around
300,000 K.
Here, we analyse part of a joint SUMER, CDS and
TRACE study, concentrating on a region which shows pro­
longed EE activity in the transition region line N v 1238 š A,
yet little evidence of such activity in another transition re­
gion line O v 629 š A observed simultaneously.
2. Observational Data
2.1. SUMER
The Solar Ultraviolet Measurements of Emitted Radiation
(SUMER) spectrograph (Wilhelm et al. 1995, Lemaire et
al. 1997) onboard Solar Heliospheric Observatory (SoHO)
was designed to provide line profile measurements at both
high spatial and spectral resolution as well as a wide spec­
tral coverage ranging from less than 500 to 1610 š A.
The data set used here was taken on 1999 June 1 start­
ing at 09:13 UT and ending at 11:01 UT. Slit 0.3 ## â120 ##
was used, exposing for 25 s on the bottom part of detector
B for Mg x 624.95 š A (in second order), O v 629.73 š A (in
second order), N v 1238.82 š A, N v 1242.80 š A, C i 1249 š A
and Si ii 1251.16 š A lines. The temporal sequence consists
of spectral windows of 120 spatial â 50 spectral pixels.
The solar rotation compensation mode was turned on for
these observations.
The O v line was observed in second order at a
wavelength of 1259.52 š A, based on Si i 1258.80 š A and
Si ii 1260.422 š A (W. Curdt, private comm.) as a chromo­
spheric reference. It is therefore blended with S ii 1259.51.
We can use two methods to estimate the strength of the
blended S ii line. From the quiet sun atlas of Curdt et al.
(2001) we can get the strengths of the two S ii lines at
1253.813 š A and 1250.587 š A, then using CHIANTI (Young
et al. 2003) one can derive the relative strength of the line
at 1259.51. Since O v was observed on the bare part of the
photocathode, the contribution from the S ii blend is in­
significant in this instance. Alternatively, SUMER allows
decomposition of blends, if lines are observed in di#erent
orders. Thus an estimate of the S ii blend in first order
can be calculated from the data set used for the QS atlas
by Curdt et al. (2001) using both the KBr and bare ex­
posures. The measured counts are linear combinations of
the contributors O v and S ii (W. Curdt, private comm.)
involving the coe#cients of well­defined sensitivity values
and photon fluxes. Solving the linear equation, yields the
photon fluxes. In the Curdt et al. (2001) quiet sun atlas,
the contribution of S ii counts was < 5% on KBr, and
therefore on bare even less.
SUMER data reduction involves several steps. We de­
compressed and reversed the raw data, applying a flat field
correction to correct the non­uniformities in the sensitiv­
ity of the detector. A geometrical distortion correction was
also applied so that the rest position of the line profiles
is on the correct spectral pixel and the slit images are
straightened. To further analyze the data set, we have used
a Gaussian­fitting procedure in the SolarSoft (SSW) 1 li­
brary to get the amplitude, central position, full width at
half maximum (FWHM) and # 2 of the line profiles. Due
to instrumental broadening, the FWHM needs to be cor­
rected by applying the con width funct 3.pro routine. The
# 2 values are used to check the goodness of the fit. The
total intensity of the lines can be derived from the peak
amplitude and FWHM.
2.2. CDS
The Coronal Diagnostic Spectometer (CDS; Harrison et
al. 1995) also onboard SoHO, was designed to obtain
spectroheliograms in a number of lines simultaneously
in the extreme ultraviolet (EUV) region of the electro­
magnetic spectrum. CDS has two spectrometers, namely,
the Grazing Incidence Spectrometer (GIS) and Normal
Incidence Spectrometer (NIS). Our data set, observed on
1999 June 1 was obtained with NIS.
Prior to and after the time sequence, 244 ## â240 ##
rasters using the 4 ## â240 ## slit were obtained. A time series
using the 2 ## â240 ## slit with a cadence of 31 s seconds was
taken from 09:01 UT to 12:13 UT. Unfortunately, the CDS
time series are o#set with respect to the SUMER point­
ing by more than ten arcseconds. Only the O v 629.73 š A
context rasters were therefore used in our study. The stan­
dard CDS data reduction routines were applied to clean
the data of cosmic rays, debias, flat­field and to correct the
o#set between NIS1 and NIS2. Details of the observations
are given in Table 1.
2.3. TRACE
Transition Region and Coronal Explorer (TRACE; Handy
et al. 1999), is a NASA Small Explorer (SMEX) mission
devoted to studying the evolution and propagation of fine­
scale magnetic fields and plasma structures throughout
the solar atmosphere. TRACE consists of a telescope with
a 30 cm primary mirror, normal incidence coatings for
three EUV bands (173, 195 and 284 š A), and interference
filters for UV bands (1216 to 1700 š A) as well as white light
(temperature ranges from # 6000 to # 2.5 â 10 6 K). The
1024 â 1024 CCD camera has a field of view of 8.5 # â 8.5 #
with a spatial resolution of 1 ## and exposure times of 0.002
to 260 s with a cadence as short as 2 s.
TRACE 173 š A images discussed here were obtained
starting at 09:02 UT and finishing at 12:00 UT on 1999
June 1. The integration time of each image was 5.8 s from
09:02 UT until 09:41 UT and 4.1 s after that time. Images
were taken every 12 s. We discuss here only the data which
were obtained simultaneously with the SUMER spectrom­
eter. In addition, white­light images were obtained every
12 min starting at 09:22 UT. The instrument was pointed
at coordinates Solar X = 422 ## and Solar Y = 401 ## .
1 http://www.lmsal.com/solarsoft/

4 Doyle et al.: Evidence for UV Explosive Event Activity Originating in the Chromosphere
Table 1. A summary of the CDS data set consists of two rasters (I and III) and a time series (II), taken on 1999 June 1.
Data Observation Start Time Exp. Time Duration Coordinates
I s16287 AR MON/v54 07:56UT 25s 55m53s (394,385)
II s16288­98 MOSS 1M/v1 09:01UT 25s 03h12m32s (401,405)
III s16299 AR MON/v54 12:14UT 25s 55m53s (423,384)
Fig. 2. Normalized light­curves for N v 1238 and O v 629 for the blue wing, red wing and main component, plus the corresponding
TRACE 173 š A band and the weak C i 1249 line. The gaps in the TRACE light curve are missing data due to the presence of
cosmic rays.
2.4. MDI
Only three full­disk MDI (Michelson Doppler
Interferometer, Scherrer et al. 1995) magnetograms
were available for the duration and pointing coordinates
of the observations. White light images were used to
co­align TRACE and MDI. The magnetograms show that
the EE appear in a region of mixed polarity weak fields.
The flux densities are close or below the noise level of
the magnetograms (12 Mx cm -1 ; see Hagenaar 2001 for
estimation method).
2.5. Alignment
Two steps were required to determine the position of the
SUMER slit in the TRACE field of view. Firstly, the
08:32 UT TRACE 173 š A image was co­registered to the
Mg ix 368 š A CDS context raster taken in the interval
07:56--08:52 UT, via cross­correlation. Similarly, that tech­
nique was used to find the match of the O v 629 š A intensity
profile along the SUMER slit (in its original form and de­
graded to the CDS spatial resolution) in one of the South­
North cross­sections of the O v CDS context raster. The
location of the SUMER slit in the TRACE images was
inferred from the comparison. Similar steps were used to
determine that the slit in the CDS time series is o#­set
with respect to the SUMER pointing.
3. Explosive event identification
As noted by many investigators, quiet­Sun profiles are
generally well represented by a single Gaussian. In an anal­
ysis of an O vi 1032 š A raster, Teriaca et al. (2004) noted
that an excellent way to locate EE's was to check the loca­
tions where one of the fitted parameters deviated by more
than 3 # from the average of its distribution. In Fig. 1, we
show a section of the O v 629 š A and N v 1238 š A time
series, where the images shown are derived from the line
width. The enhanced regions are those with larger widths,
i.e. those that we assigned a preliminary flag as being pos­
sible EE's. The individual profiles were then checked to see
whether there was evidence of a high velocity explosive
component.

Doyle et al.: Evidence for UV Explosive Event Activity Originating in the Chromosphere 5
Fig. 3. The O v and N v line profile at 09:53 UT, i.e. event (b).
The quiet Sun profile is over­plotted for comparison (dashed
line).
4. Results
For illustrative purposes, we show five temporal loca­
tions, indicated by the letters (a), (b), (c), (d) and (e)
for summed pixels 47 to 50 along the slit direction. In the
N v profiles, these times indicated EE activity, with sev­
eral of them indicating mass flows in excess of 150 km s -1
and line intensity increases sometimes exceeding the nor­
mal `quiet' Sun component. With exception of event (a),
the O v line profiles fail to show similar evidence of large­
scale EE activity at some locations. Instead, we see either
none or in some instances, only a minor indication of mass
flows (see Fig. 2). Here we show light­curves for the three
components of both lines: blue, main and red. The main
component results from summing over 9 spectral pixels,
i.e. ± 45 km s -1 for N v and half this value for O v, with
the blue­wing and red­wing components extending from
these values out to # 160 km s -1 and excluding the chro­
mospheric Si ii line to the red of the O v. In addition, the
corresponding TRACE 173 š A light­curve is given, plus the
weak chromospheric C i 1249 line.
It is at first hard to understand the lack of similar ac­
tivity in both lines as they are considered to be formed
at a similar temperature. For example, using CHIANTI,
the peak formation temperature of N v is 200,000 K while
O v has a peak formation temperature of 250,000 K, con­
sidering the ionization balance calculations of Mazzotta et
al. (1998).
Several other EE's were detected during the # 100
minutes of observations at di#erent locations along the
slit, e.g. between pixels 57 and 60 along the slit around
09:35 UT. Here, both the N v and O v lines show the
presence of prolonged EE activity. The largest of the N v
EE's is at location (b), with the event showing a peak
counts of close to 500 cts/pixel in the blue­shifted plasma
compared to the quiet Sun counts of 130 cts/pixel in the
stationary component. This event is shown in more detail
in Fig. 3. The peak count rate in the O v line is similar in
both the quiet sun and EE (b).
5. Discussion
In an e#ort to try to understand how the EE's are eas­
ier identified in the N v line, while they are practically
absent in O v despite the fact that both of these lines
are formed at very similar temperatures, we look closely
at their formation. The usual practice when considering
atomic processes in high­temperature, low­density plas­
mas, such as discussed here, is to adopt the coronal ap­
proximation. This treats the populations of excited states
of ions via an excitation balance of collisional excitation,
usually from the ground state, by electrons, and radia­
tive decay. The ionization state is established as a balance
of electron impact ionization from the ground state and
radiative plus dielectronic recombination. In the simplest
version of such modeling, secondary collisions with excited
states are neglected. However, there is no consistent treat­
ment of metastable states with populations comparable
to the ground state. Thus, in the coronal limit, ionization
and excitation balance are independent of electron den­
sity. Recent work by Doyle et al. (2005) has looked at
this problem regarding formation of the Li­like lines only.
Here, we have a line from a Li­like ion and the other is
from a Be­like ion.
The generalized collisional­radiative picture (Summers
& Hooper, 1983), which builds on the collisional­radiative
theory of Bates, Kingston & McWhirter (1962), allows
such an analysis. In detail, the collisional ionization and
redistribution processes from excited states, ignored in the
coronal model, are included. The populated metastable
states are determined via an elaborated ionization bal­
ance along with the ground states. These were computed
within the Atomic Data and Analysis Structure (ADAS;
Summers 2004 2 ) framework, which is a collection of fun­
damental and derived atomic data, and codes that ma­
nipulate them. The data are organized in a form that al­
lows generation of the collisional­radiative matrix, which
is manipulated to form the derived e#ective recombina­
tion and ionization coe#cients. There is no significantly
populated metastable level in the case of Li­like ions, but
2 http://adas.phys.strath.ac.uk

6 Doyle et al.: Evidence for UV Explosive Event Activity Originating in the Chromosphere
the metastable resolved picture matters when considering
recombination and ionization to and from the He­like and
Be­like stages. Further details are given in Doyle et al.
(2005).
In Fig. 4 we show the density dependent contribution
function for N v 1238 and O v 629 for four values of
the electron density; 10 6 cm -3 representing the low den­
sity limit, 10 9 cm -3 for a typical quiet Sun electron den­
sity, plus 10 11 cm -3 and 10 12 cm -3 for an active region.
Here we see that with increasing electron density, both
Fig. 4. The density dependent contribution function for
O v 629 š A and N v 1238 š A for four values of the electron
density, Ne = 10 6 cm -3 , 10 9 cm -3 , 10 11 cm -3 and 10 12 cm -3 .
lines shift to slightly lower temperatures. However, the
major di#erence is in the relative increases in the line flux
with increasing density. For N v, increasing the density to
10 11 cm -3 results in a 60% increase in the line flux, while
O v shows a 30% decrease. Increasing the electron density
to 10 12 cm -3 results in a factor of two decrease in the
O v flux. This decrease in the line flux can be explained
by considering the relative populations of the metastable
and ground terms of the O +4 ion. At low densities the
metastable population is negligible but, on increasing the
density, the population becomes significant and acts as a
sink to higher levels, thus causing the 629 š A transition to
be depleted.
Hence, a possible explanation for the lack of major ac­
tivity in the O v line for events (b), (d) and (e) could be
that these EE's occur in the lower chromosphere, there­
after in a high electron density region. Since EE's are pre­
dominately line­shift events as opposed to line flux en­
hanced events, the increased sensitivity of N v by a factor
of two compared to O v which is decreased by a factor
two, means that these EE's are less visible in O v.
This is consistent with the findings of Madjarska &
Doyle (2002) who found time delays between the chro­
mospheric and transition region lines in some EE's using
high cadence observations (10 s exposure time) obtained
with the SUMER spectrometer in H i Ly 6 (20 000 K) and
S vi (200 000 K). This suggested that the process which
generates these jets first occurs in the low chromosphere
although they are best observed in transition region lines.
In another study of EE's, Doyle et al. (2003) used
a non­Maxwellian electron distribution to explain how
``coronal'' plasma as detected in the TRACE imager with
the 173 š A filter could derive from temperatures around
#300,000K, as opposed to #800,000K. These authors had
an example of a bi­directional jet registered in the chro­
mospheric and the transition region lines but not showing
any detectable signature in the coronal line, yet was clearly
detected in the TRACE 173 š A pass­band. Hence, it is pos­
sible that these transition region lines could be e#ected
by a non­Maxwellian electron distribution, thus emitting
at close to chromospheric temperatures. However, we rule
this out as a possible explanation for the lack of activity
in events (b), (d) and (e), although this may not be the
case for event (c). In Fig. 5 we show the N v and O v line
profiles, in addition to an expanded plot of the various
light­curves for this event.
Fig. 2 shows that this event is observable in both O v
and N v, in addition to TRACE, thus perhaps questioning
the increased electron density suggestion. However, a close
inspection of Fig. 1 shows that part of this event is shifted
a few arcsecs from the other events and thus is likely to be
occurring in di#erent plasma conditions than the others.
Further evidence for the chromospheric nature of
events (b), (d) and (e) comes from the observed increase
in the Si ii 1251 line to the red of O v, see Figs. 1, 3 &
5. In this low temperature line, although we only observe
the red­wing, a large flux increase is apparent. However,
for event (c) (see Fig. 5) the Si ii line has a similar inten­
sity in both the quiet Sun and during the EE. This would
suggest that unlike the other events, this EE is produced
higher in the atmosphere, probably in the transition re­
gion at around a temperature of 250,000 K. Furthermore,
as shown in Fig. 2, the C i line shows an increase in inten­
sity for events, (a), (b), (d) and (e), consistent with the
idea that these EE's are produced in a low temperature
plasma, but no apparent increase is seen for event (c). For
event (c), the N v line shows very broad profiles over two
consecutive exposures with only minimum evidence for
secondary components. The O v on the­other­hand shows
a secondary blue­shifted component, with broadened pro­
files stretching up to 5 ## to the north. The TRACE 173
light­curve peaking some 40 s after the rise in N v and
O v may very well be coronal in origin. To test this idea
further, we looked closely at the Mg x 625 in second order
and the weak C i 1249 line.
The Mg x 625 š A line is observed as a second order line
at 1249.90 š A, therefore, it overlaps with several first order
chromospheric lines, namely P ii 1249.82, Mg ii 1249.93
and Si ii 1250.09. A detailed discussion on these blends is
given by Teriaca et al. (2001) and Doyle et al. (2003). The
latter authors performed a detailed spectroscopic analy­
sis for the largest of these EE's (i.e. event b) in order to
estimate the contribution of the spectral lines blending

Doyle et al.: Evidence for UV Explosive Event Activity Originating in the Chromosphere 7
Fig. 5. Left: the O v and N v line profiles at 10:00 UT, i.e. event (c). The quiet Sun profile is over­plotted for comparison
(dashed line). Right: close­up of the time series shown in Fig. 2 corresponding to the interval 45--50 minutes.
to Mg x when recorded on the KBr part of detector B.
The average contribution for the Si ii 1250.09 š A line was
found to be between 35 and 40 %. Another few percent can
come from the Mg ii and the P ii lines leading to a total
first­order contribution of #40%. The result was that the
small increase observed in Mg x was due to the chromo­
spheric contribution and thus the small observed increase
in TRACE 173 did not come from the corona. These au­
thors suggested the presence of a non­Maxwellian elec­
tron distribution to explain the TRACE data. Although
in event (c) we have a much larger intensity increase in the
TRACE 173 passband, as in previous studies by Teriaca
et al. (2001) and Doyle et al. (2003), there is no appar­
ent increase in the Mg x line, suggesting either that the
TRACE 173 increase is not from the coronal region or
that the EE jet manifests itself as a flux increase at coro­
nal temperatures as opposed to a velocity feature.
The above findings therefore suggest two types of
EE's; one formed in the low chromosphere and the other
formed in the mid­to­high transition region. This there­
fore suggests that single temperature line profile informa­
tion is not su#cient to clarify the nature of these events.
Furthermore, this again shows the importance of consid­
ering a full atomic model as opposed to the commonly
adopted assumption of ionization and recombination to
the ground­state only.
Acknowledgements. We would like to thank the SUMER team
at Max Planck Institute for Solar System Research (MPS), and
CDS and TRACE teams at Goddard Space Flight Center for
their help in obtaining the present data. SUMER and CDS
are part of SoHO, the Solar and Heliospheric Observatory,
a mission of international cooperation between ESA and
NASA. Research at Armagh Observatory is grant­aided by the
N. Ireland Dept. of Culture, Arts and Leisure. BI wishes to
thank E. O'Shea for productive discussions, plus M. Madjarska
and L. Xia for their help with the SUMER software, and
W. Curdt for discussions on the wavelength calibration and
line blending. This work was supported in part by a PRTLI
research grant for Grid­enabled Computational Physics of
Natural Phenomena (Cosmogrid). CHIANTI is a collabora­
tive project involving the NRL (USA), RAL (UK), and the
Universities of Florence (Italy) and Cambridge (UK)
References
Bates, D.R., Kingston, A.E. & McWhirter, R.W.P., 1962, Proc.
Roy. Soc., A267, 297
Brueckner, G.E. & Bartoe, J.­D.F., 1983, ApJ 272, 329
Chae, J., Wang, H., Lee, C. ­Y., Goode, P. R., & Sch˜uhle, U.,
1998, ApJ, 497, L109

8 Doyle et al.: Evidence for UV Explosive Event Activity Originating in the Chromosphere
Curdt, W., Brekke, P., Feldman, U., Wilhelm, K., Dwivedi,
B.N., Sch˜uhle, U. & Lemaire, P., 2001, A&A 375, 591
Dere, K.P., 1994, Adv. Space Res. 14, 13
Doyle, J.G., Dzif’c’akov’a, E. & Madjarska, M.S., 2003, Sol. Phys.
218, 79
Doyle, J.G., Summers, H.P. & Bryans, P., 2005, A&A 430, L29
Hagenaar, H.J., 2001, ApJ 555, 448
Handy, B. N., Acton, L. W., Kankelborg, C. C., et al., 1999,
Sol. Phys., 187, 229
Harrison, R. A., Sawyer, E. C., Carter, M. K., et al., 1995, Sol.
Phys., 162, 233
Madjarska, M.S. & Doyle, J.G., 2002, A&A 382, 319
Mazzotta, P., Mazzitelli, G., Colafranceso, S. & Vittorio, N.,
1998, A&A S 133, 403
Lemaire, P., Wilhelm, K., Curdt, W., et al., 1997, Sol. Phys.,
170, 105
Scherrer, P. H., Bogart, R. S., Bush, R. I., et al., 1995, Sol.
Phys., 162, 129
Summers, H.P. & Hooper, M.B., 1983, Plasma Physics, 25,
1311
Teriaca, L., Madjarska, M.S. & Doyle, J.G., 2001, Sol. Phys.
200, 91
Teriaca, L., Banerjee, D., Falchi, A., Doyle, J.G. & Madjarska,
M.S., 2004, A&A 427, 1065
Wilhelm, K., Curdt, W., Marsch, E., et al., 1995, Sol. Phys.,
162, 189
Young, P.R., Del Zanna, G., Landi, E., Dere, K.P., Mason,
H.E. & Landini, M., 2003, ApJS 144, 135