Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://star.arm.ac.uk/~ambn/324jgd.ps Äàòà èçìåíåíèÿ: Wed Dec 8 12:36:34 1999 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 21:00:49 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: coronal hole

Polar Plumes and Inter­plume regions as observed by
SUMER on SOHO
D. Banerjee, L. Teriaca & J.G. Doyle
Armagh Observatory, Armagh BT61 9DG, N. Ireland.
dipu@star.arm.ac.uk, lte@star.arm.ac.uk, jgd@star.arm.ac.uk
P. Lemaire
Insitut d'Astrophysique Spatiale, Unite Mixte CNRS, Universite Paris XI, Bat
121, F­91405, Orsay, France
(Received ..... ; Accepted in final form .....)
Abstract. We present observations of O vi 1032 š A line profiles obtained with the
SUMER instrument on SOHO extending from the solar disk to 1.5 R fi above the
limb in the north polar coronal hole. Variations of the intensity and line width in the
polar plume and inter­plume regions are investigated. We find an anti­correlation
between the intensity and the line width in the plume and inter­plume regions with
detailed plume structures been seen out to 1.5 R fi . Possible implications regarding
the magnetic topologies of these two regions and related heating mechanisms are
discussed. The O vi line width measurements are combined with UVCS output to
provide an overview of its variations with height extending up to 3.5 R fi . We find
a linear increase of the line width from 1 to 1.2 R fi , then a plateau followed by a
sharp increase around 1.5 R fi .
1. Introduction
With the launch of SOHO, and in particular the onboard SUMER
instrument (Wilhelm et al. 1997a), line width measurements of coronal
lines as a function of position above the limb is once again possible.
Wilhelm et al. (1997b) and Lemaire et al. (1997) has observed emis­
sion lines for a large number of ions from the solar disk out to 1.6 R fi .
Recently Doyle, Banerjee & Perez (1998) showed an increase of the line
width up to ¸ 70,000 km above the limb for quiet Sun conditions at the
equatorial western limb. From Si viii line pair observations in a polar
coronal hole, Banerjee et al. (1998) found evidence for the presence of
Alf'ven waves. They have also shown that these waves can carry signif­
icant energy flux through the transition region and into the corona. In
this paper the widths of the O vi line will be used to determine the
thermal and turbulent speed associated with the solar coronal plasma.
The observations were made by the Solar Ultraviolet Measurements
of Emitted Radiation (SUMER) spectrometer on the Solar and Helio­
spheric Observatory (SOHO) in the north polar coronal hole of the Sun.
If the velocity distribution can be assumed as a Gaussian, the result­
ing Doppler broadening (or half width of the line at 1/e of the peak

2 D. Banerjee
intensity) is given by,
\Delta– D = –v 1=e =c (1)
Here – is the wavelength and c the velocity of light. Within about five
Doppler widths of line center, a Gaussian function represents the pro­
file well. Gaussian velocity distribution assumes a Maxwellian velocity
distribution (random turbulent motion), which is probably the domi­
nant factor in line broadening in the transition region where unresolved
velocities in all directions (in the integration column along the line of
sight) may contribute. Other broadening, such as collisional, may be
small and will produce some Lorentzian shape. For optically thin UV
emission lines, there is little need to consider the more complicated form
of the profile (Mariska 1992). Moreover, with almost no exceptions we
have seen that our line profiles are Gaussian in shape. Now the line
of sight speed (v 1=e ) distribution includes two distinct contributions,
namely the thermal motions and the small scale unresolved turbulent
motions. If the corresponding velocity distributions are both assumed
to be Gaussian, we can superpose these distributions and simply add
their variances such that,
v 2
1=e = 2kB T eff
M
= 2kB T i
M
+ ¸ 2 (2)
where ¸ is the non­thermal speed, related to the wave amplitude by
¸ 2 = 1
2 ! ffi v 2 ?, where the factor of 2 accounts for the polarization
and direction of propagation of a wave relative to the line of sight (for
Alfv'enic wave types), T i the ion temperature, and M the ion mass. One
has to discriminate between the two terms on the right hand side of the
equation in order to get a physical insight to the heating mechanisms
in the corona from the observational data. Observations have revealed
that plasma conditions in polar plumes are quite different from inter­
plume lanes (Hassler et al. 1997; Wilhelm et al. 1998). In this paper we
will concentrate on the plume and inter­plume regions separately. We
will study the variations of the intensity and the inferred v 1=e from the
observed FWHM. This will provide clues to the structure of these two
regions and the possible role of magnetic fields.
2. Observations & the Data Reduction
SUMER is a normal incidence spectrograph operating over the wave­
length range ¸500 š A to 1600 š A. The off­axis parabola mirror is move­
able in two dimensions around the focal point allowing pointing of the
instrument independent of the spacecraft pointing. Four slits are avail­
able; 4 00 \Theta 300 00 , 1 00 \Theta 300 00 , 1 00 \Theta 120 00 and 0:3 00 \Theta 120 00 . The observations
sumer—plume—v4.tex; 16/11/1999; 11:49; no v.; p.2

Polar Plumes and Inter­plume regions 3
Figure 1. The locations of the SUMER datasets in the north coronal hole super­
imposed on an EIT image of He ii 304 š A obtained on 3rd June 1996 at 17:51 UT
(courtesy of the EIT consortium).
reported here were obtained on 3 rd June 1996 starting at 11:01 UT
and ended at 15:00 UT. Detector A (Siegmund et al. 1994) was used
for this study. The observational sequence reported here consisted of
three parts, all regarding the North Polar Coronal Hole (NPCH). The
first part of the sequence consists of a raster scan obtained using the
0:3 00 \Theta 120 00 slit with an exposure time of 16 sec. The scan started at
x = \Gamma152 arcsec; y = 955 arcsec (as measured from disk center) with
the slit aligned in the north­south direction. After each exposure, the
slit was moved by 3:01 arcsec, accumulating an image covering an area
of ¸ 300 00 \Theta 120 00 with a resolution of 3 00 \Theta 1 00 . Fig. 1 shows an image of
the NPCH region taken with EIT (Moses et al. 1997) in He ii 304 š A at
17:51 UT on June 3, 1996 with the spectroheliogram superimposed. It
can be clearly seen from the figure that the north coronal hole during
the solar minimum phase was extended onto the disk.
In order to increase the signal to noise ratio we binned the original
image by 2 \Theta 2 spatial pixels in the x and y directions respectively with
a final resolution of 6 00 \Theta 2 00 . Only the lower 90 arcsec of the image were
sumer—plume—v4.tex; 16/11/1999; 11:49; no v.; p.3

4 D. Banerjee
considered (see Figs. 2 and 3 lowest panels). The second and third part
of the sequence was intended to obtain high signal to noise line profiles
in the plume and inter plume regions above the limb out to 1.5 R fi .
For the second scan, an exposure time of 60 sec and the 4 00 \Theta 300 00 slit
was used. Scans started at x = \Gamma143 arcsec; y = 1150 arcsec and, after
each exposure, the slit was moved by 3:76 arcsec, covering an area of
¸ 278 00 \Theta 300 00 arcsec 2 with a resolution of 3:8 00 \Theta 1 00 . In this case we
binned over 15 pixels in the y direction, obtaining a final resolution of
3:8 00 \Theta 15 00 . Only the lower part (with higher S/N ratio) of the image
was used (see Figs. 2 and 3).
Finally, for the third scan exposure times of 150 sec and the 4 00 \Theta 300 00
slit was used. Scans started at x = \Gamma149 arcsec; y = 1300 arcsec and,
after each exposure, the slit was moved by 5:65 arcsec, covering an area
of ¸ 287 00 \Theta 300 00 . In this case we binned over 20 pixels in the y direction
(see Figs. 2 and 3).
For the SUMER instrument, the process of data reduction involves
three main steps; flat field subtraction, de­stretching, radiometric cali­
bration and slit effects correction. All these corrections were done using
various IDL routines from within the SUMER software tree. The mea­
surement of line parameters were performed using the Genetic Algo­
rithm (GA) of Charbonneau (1995). A complete examination of the
reliability of GA with respect to other algorithms was performed by
McIntosh et al. (1998). An estimation of the errors in the derived
parameters was obtained using the GA­derived parameters as input for
a ``classical'' IDL­CURVEFIT procedure (Peter, 1999; Peter & Judge,
1999 and Teriaca et al. 1999). A total of 4403 fits were hence performed.
3. Results
We first approximate the line shapes by Gaussian fits in order to deter­
mine the most probable speed along the line of sight. The instrument
line width correction was then applied to obtain the Doppler width
\Delta– D = FWHM=(2
p
ln2), where FWHM is the full width at half
maximum of the line after instrumental effects have been taken out. The
measured line profiles will also depend on the instrumental character­
istics. In addition, the slit width has to be considered. The latter effect
can not be treated as a Gaussian distribution and a de­convolution
is required. This has been done using the standard SUMER software.
The magnitude of this correction is small, although it does result in
a correction in the derived velocity of ¸2.5 km s \Gamma1 for the 4 00 \Theta 300 00
slit. Equation (1) was then used to calculate the most probable speed,
v 1=e . We do not attempt to separate the line widths into a contribution
sumer—plume—v4.tex; 16/11/1999; 11:49; no v.; p.4

Polar Plumes and Inter­plume regions 5
1
Fig. 1.
Figure 2. Intensity map of the north polar coronal hole as seen in O vi on June
3rd 1996. Strong plumes and inter plume lanes can be identified in this diagram
out to 1.6 R fi above the limb. The lower panel represents the on­disk scan. The
east­west scan extends from \Gamma150 to +150 arcsec. For the upper panel the contrast
was enhanced by normalizing to the average intensity profile in the y­direction. The
strongest plume and inter­plume lane is traced out by a broken and continuous line
respectively.
sumer—plume—v4.tex; 16/11/1999; 11:49; no v.; p.5

6 D. Banerjee
1
Fig. 1.
Figure 3. The half width of the O vi line at 1/e of the peak intensity in km s \Gamma1
for the same region as in Figure 2.
sumer—plume—v4.tex; 16/11/1999; 11:49; no v.; p.6

Polar Plumes and Inter­plume regions 7
Figure 4. A plot of v 1=e
(solid line) and intensity (dashed line) at the limb using
the on disk scan data.
from the thermal part, T i and the non­thermal motions, ¸, as the data
available do not allow to do this. Figure 2 shows the contrast enhanced
total line intensity of the O vi 1032 š A line over a north polar coronal
hole.
Particular attention needs to be paid to the stray light contribution
to the total line intensity as previously done by Hassler et al. (1997).
They suggested that the contribution of stray light to O vi profiles
approaches 20% at 1.2 R fi and no coronal contribution to the O vi
profile is present at 1.6 R fi . The level of scattered light seems to be
independent of the azimuthal position around the solar disk (Lemaire
et al., 1997). This allows us to proceed in two steps. First we will
study the plume­interplume behaviour at different heights above the
limb without subtracting any scattered light contribution. Then we
will estimate (and correct for) the scattered light contribution along
two different locations away from the solar disk.
In Figure 3 we produce a velocity image (assuming a single Gaussian
profile) of the entire scanned region on 3rd June 1996. To look for
possible correlation/anti­correlation between the line intensity and the
most probable speed, we concentrate on specific heights. Using the disk
scan we plot the line intensity (dashed line) and the most probable
velocity (solid line) along the limb in Figure 4. In Figure 5 we plot the
variation of intensity with height (lower panel) along a radial path, from
Sun center to outside the limb using the disk data. In the lower panel
sumer—plume—v4.tex; 16/11/1999; 11:49; no v.; p.7

8 D. Banerjee
Figure 5. Variation of the most probable speed v 1=e
(upper panel) and intensity
(lower panel) with height using the on­disk scan data. The solid line corresponds to
a typical inter­plume region and the dashed line corresponds to a plume as seen on
the limb in Fig. 2. Velocities in the upper panel correspond to the dashed line. The
vertical dotted line represents the rough location of the solar limb.
of Figure 5 the solid line traces one typical inter­plume lane from the
disk to outside the limb and the dashed line traces one typical plume
structure (see Figs. 2 & 3 for locations). The upper panel of Figure 5
shows the most probable speeds with error bars at different heights
corresponding to the plume. The rough location of the solar limb is
marked by a vertical dotted line. The sharp peak near the limb of the
dashed line corresponds to the brightest plume in Figure 2. We find a
correlation between line intensity and width. The data suggests that
the network structures as visible on the disk continue as plumes outside
the disk (see Fig. 2). Correlation of line width with enhanced emission
in the network has already been published in the literature (Warren et
al. 1997). We could clearly see on the disk part of Figure 5 that the
enhanced intensity of the dashed line leads to additional velocities in
the upper panel. Furthermore to quantify the correlation between the
line intensity and the width, we have plotted several scatter plots. As
an example we show a scatter plot of the disk observation in Figure 6.
sumer—plume—v4.tex; 16/11/1999; 11:49; no v.; p.8

Polar Plumes and Inter­plume regions 9
Figure 6. Scatter plot of the most probable speed v 1=e vs. intensity for the disk
part.
To produce this scatter plot we have selected a region of the disk scan
within the limb. Though there is a considerable scatter, it does suggest
that brighter regions have higher average most probable speeds than
faint regions.
As we go off the limb, Figure 7 shows the line intensity (dashed line)
and the most probable velocity (solid line) along three strips tangent
to the limb at 1.07, 1.12 and 1.22 solar radii (using the off limb scan).
We find clear evidence of an anti­correlation between the intensity and
v 1=e . If we go out far from the limb, the lowest panel of Figure 7 shows
that at 1.22 R fi the anti­correlation is rather weak, though we do have
anti­correlation at locations x=\Gamma50, 70 & 120. Also note that at this
height above the limb the plume structures have expanded slightly non
radially (see Fig. 2). The left­most side of the plume seems to have some
angle (and curvature) with the radial polar axis, which could imply that
a double plume structure becomes visible at some distance away from
the solar limb. This observed anti­correlation between intensity and
width in polar plumes have been reported by Antonucci (1997) and
Noci et al. (1997) with the UVCS instrument and also by Hassler et al.
(1997) with SUMER.
sumer—plume—v4.tex; 16/11/1999; 11:49; no v.; p.9

10 D. Banerjee
Figure 7. A plot of the most probable speed v 1=e
(solid line) and intensity (dashed
line) along three strips tangent to the limb at 1.07, 1.12 and 1.22 solar radii to
illustrate the anti­correlation between intensity and velocity.
sumer—plume—v4.tex; 16/11/1999; 11:49; no v.; p.10

Polar Plumes and Inter­plume regions 11
Figure 8. Variation of the intensity with height in a north polar coronal hole along
a polar plume (solid line) and inter plume regions (dashed line) for O vi 1037 and
C ii 1037.
3.1. Stray light correction
Despite the very high quality of the telescope mirror, the level of scat­
tered light is not negligible when observations are carried out in faint
region such as Polar Coronal Holes (Lemaire et al., 1997; David et al.,
1998). An estimation of the stray contribution to the O vi 1032 line
profile can be made using a purely chromospheric line such as C ii 1037
(Hassler et. al., 1997). This line does not have any coronal contribu­
tion and is, hence, entirely due to stray light. In Figure 8 we show the
behaviour of the intensity with height above the limb for O vi 1037 and
C ii 1037. Note the plume­interplume contrast in O vi. If C ii emis­
sion is due only to scattered light we would not expect any difference
between plume and interplume (Hassler et al., 1997). The difference
noted in Figure 8 near the disk may be due to the difficulty of fitting
the C ii line particularly when it is only a tiny fraction of the nearby
O vi 1037 line. Furthermore a proper fit of the C ii 1037 line is more
difficult because the spectra nearest to the disk (up to ¸ 1.2 R fi ) are
only 50 pixel wide. These problems lead, essentially to an overestima­
tion of the line width of C ii near the limb. In fact the peak intensity
for C ii show the same values in plume and interplume. Data relative to
interplume were, hence, used for the calculation of the stray light con­
tribution. We assumed that there is no coronal contribution above 1.6
solar radii (Lemaire et al., 1997; Hassler et al., 1997). The stray light
sumer—plume—v4.tex; 16/11/1999; 11:49; no v.; p.11

12 D. Banerjee
Figure 9. Variation of the most probable speed v 1=e and intensity with height in a
north polar coronal hole along a polar plume (dashed line) and inter plume regions
(solid line). The diamonds correspond to inter plume and the square boxes corre­
spond to a plume. The dotted line (in the lower panel) represents the chromospheric
line C ii and its contribution is entirely due to stray light.
profile in O vi 1032 was obtained using observations at 1.8 solar radii
of March 1997. Differences coming from the use of a different detector
(B) and a different slit (1\Theta300) were taken into account. These obser­
vations were planned as part of the program `The Sun as a star' (P.
Lemaire). The profile at 1.8 R fi was then normalized according to the
stray light trend inferred from C ii.
In Figure 9 we trace out the plume (dashed line) and inter­plume
(solid line) regions showing the variation of velocity (upper panel)
and intensity (lower panel) with height. One can clearly see additional
speeds in the inter plume region which tends to disappear at 1.35 R fi .
This may be due to the fact that the left plume have some curvature
with the polar axis (see Fig. 2) and thus the dashed line at around 1.3
R fi does not follow exactly the maximum intensity and is at the edge
of the expanded plume.
The present line width measurements can be combined with UVCS
output to provide an overview of its variation with height out to 3.5 R fi .
sumer—plume—v4.tex; 16/11/1999; 11:49; no v.; p.12

Polar Plumes and Inter­plume regions 13
Cranmer et al. (1999a) have presented a comprehensive and self con­
sistent empirical model of several plasma parameters in the extended
corona above a polar coronal hole. The model is derived from obser­
vations with UVCS. Their model calculations show that the velocities,
which are valid between 1.5 and 3.5 R fi can be approximated by the
best­fit function of the form,
v 1=e = 74:3 + 200
Ÿ
0:23
i
R fi
r
j 0:0327
+ 4:69
i
R fi
r
j 2:36
+1:58 \Theta 10 5
i
R fi
r
j 22:9
– \Gamma1
(3)
Combining results from SUMER and UVCS, Fig. 10 shows the varia­
tion of v 1=e out to 3.5 R fi . The diamonds represent our observations
with O vi. The star symbols represent the observed values between 1.5
and 3.5 R fi from O vi profiles of UVCS (Kohl et al. 1998). The dashed
line represents the best­fit function (Eq. [3]) as derived from the empir­
ical model of Cranmer et al. (1999a). It is clearly seen that a simple
extrapolation of equation (3) down to 1 R fi does not fit the SUMER
data. Based on our observations and UVCS analysis of Cranmer et al.
(1999a) we propose a polynomial equation between 1 and 3.5 R fi as,
v 1=e = 55
i
r
R fi
j
+ 150
Ÿ
0:23
i
R fi
r
j 0:0327
+ 4:69
i
R fi
r
j 2:36
+1:58 \Theta 10 5
i
R fi
r
j 22:9
– \Gamma1
(4)
In Figure 10 the solid line represents the polynomial equation (4).
An enlarged view of the SUMER data is further presented as an inset.
The Si viii results from Doyle et al. (1999) have also been over­plotted
(represented by triangles) in the inset.
4. Discussion
SUMER line profile observations of O vi have been published by Wil­
helm et al. (1998) and Hassler et al. (1997). We reconfirm their main
result that the line profiles are slightly wider in inter­plume lanes than
in plumes. This effect is significant in determining the structure of these
two regions and the possible role of magnetohydrodynamic waves in
heating and accelerating the solar wind. Our results also provide a mor­
phological view of the north polar coronal hole. Based on simultaneous
data from five instruments on board SOHO, Deforest et al. (1997) have
presented a morphology of the plumes in the south polar region. They
sumer—plume—v4.tex; 16/11/1999; 11:49; no v.; p.13

14 D. Banerjee
Figure 10. Variation of the most probable speed v 1=e with height in a north polar
coronal hole. The diamonds represent our results from O vi for the inter­plume
region, the star symbols correspond to the UVCS results taken from Kohl et al.
(1998) and the triangles correspond to Si viii (Doyle et al. 1999). The solid line
represents the polynomial equation given by Eq. (4) and the dashed line represents
the best­fit function as given by Eq. (3) due to Cranmer et al. (1999a).
show that the plumes arise from small (¸ 2­5 arcsec diameter) quies­
cent, unipolar magnetic flux concentrations, on chromospheric network
cell boundaries. They are also cooler and denser than the surrounding
coronal hole. Our on disk results do indicate that the plume structures
are probably extended from the network boundaries.
The anti­correlation between the intensity and the velocity v 1=e can
be attributed to a lower electron temperature of the plumes as com­
pared to the inter­plume lanes observed in Mg ix (706/750 š A) and
Si viii (1440/1445 š A) with SUMER (Wilhelm et al. 1998). More recent
results from UVCS (Cranmer et al. 1999a) also shows that the elec­
trons are much cooler than the ions. But at the same time it should
be pointed out that most recent studies show that the assumption of
collisional ionization equilibrium and the common notion that T e = T i
in the coronal hole plasma can not be made any more (Wilhelm et al
1998; Doyle et al. 1999). Tu et al. (1998) have pointed out that in a
sumer—plume—v4.tex; 16/11/1999; 11:49; no v.; p.14

Polar Plumes and Inter­plume regions 15
coronal region a few tens of arc secs above the limb the ion kinetic tem­
peratures may not have anything to do with the electron temperature
or the formation temperature.
The electron temperature estimates do not provide a definitive clue
on the ion temperature, thus it is not conclusive from the SUMER
observations that the ions in the inter­plumes are hotter than those
in the plumes. Furthermore a closer inspection of the inset of Fig. 10
(both are for inter­plume lanes) reveals that v 1=e from O vi is higher
than Si viii. Since the turbulent line broadening should be the same for
different ions, the difference at a particular height should then be due
to the thermal contribution. Thus we also found like Tu et al. (1998),
that the ion thermal speed decreases with increasing mass per charge.
The other possibility is that the non­thermal component of equation
(2) is responsible for the anti­correlation. If the thermal contribution in
the plume and inter­plume is roughly the same then the excess broad­
ening in inter­plume lanes proves the existence of additional waves or
turbulence. Recent studies (Koutchmy et al. 1997; Wilhelm et al 1998)
have suggested that the source of the fast solar wind lies in the low
density inter­plume lanes and not in the plumes. Our excess broaden­
ing and anti­correlation of intensity and velocity can be explained in
terms of Alfv'en waves in the inter­plume lanes, which are believed to
be the acceleration site of the fast solar wind. Furthermore, the Alfv'en
waves with an amplitude of 30­50 km s \Gamma1 (as observed by Banerjee et
al. 1998) at the base of the coronal hole can generate non­linear solitary
type of wave, which can contribute significantly to solar wind acceler­
ation in open magnetic field structures. Cranmer et al. (1999b) have
presented theoretical models of the dissipation of the high frequency
(10 to 10,000 Hz) ion cyclotron resonant Alfv'en waves and also suggest­
ed that these waves can be generated gradually throughout the wind
rather than propagating up from the base of the corona. They have
further shown that these high frequency ion cyclotron resonant Alfv'en
waves can heat and accelerate ions differently depending on their charge
and mass which may explain the difference in observed v 1=e of Si viii
and O vi as shown in the inset of Figure 10.
Now we turn our attention to the combined results of SUMER and
UVCS and concentrate on the radial dependence of the line width from
limb, extending up to 3.5 R fi . This allows us to investigate physical
conditions of the inner solar corona and corona­heliospheric interface.
Our results will provide precise input parameters for solar wind models.
Comparing different observations from UVCS, Dodero et al. (1998)
have reported a high degree of anisotropy in the oxygen ion velocities.
UVCS observations indicate the presence of motions only perpendicular
to magnetic field (Antonucci et al. 1997). The fast solar wind in the
sumer—plume—v4.tex; 16/11/1999; 11:49; no v.; p.15

16 D. Banerjee
polar regions undergoes a large acceleration in the region close to the
Sun and at 3 R fi a velocity between 350 km s \Gamma1 and 500 km s \Gamma1 is
attained (see Fig. 10). If the excess broadening is due to the propagation
of Alfv'en waves then Figure 10 further suggests an effective breakdown
of the WKB approximation for the propagation of linear Alfv'en waves
around 1.25 R fi . If the linear theory would have been correct then one
would expect a linear increase of velocity with height, but we rather
notice a plateau (see Fig. 10) around 1.4 R fi (see also Doyle et al.
1999) followed by a sharp increase at 1.5 R fi , which suggests that non­
linearity becomes important around 1.25 R fi . We should point out here
that the Si viii line is a coronal line and thus is not affected by scattered
light. The inset of Figure 10 does indicate the presence of this plateau
very clearly, as compared to the O vi line, where the values could be
slightly overestimated because of the assumed stray light profile.
Furthermore one should note from Figure 10 that relatively sharp
variations occurring around 1.5 R fi may indicate the location where the
thermalization and the isotropization times of various species begins to
exceed the local coronal expansion time. Esser et al. (1999) from a
study of Mgx and O vi lines (observed with UVCS) found a transition
from collisional to collisionless plasma between 1.75 to 2.1 R fi in a polar
coronal hole. Cranmer et al. (1999a) also indicates the presence of this
transition. For line width measurements, UVCS data are not available
below 1.5 R fi , so the combined SUMER and UVCS datasets allows us
to locate this transition point with better precision. It is interesting to
note that Doyle et al. (1999) have shown that the radial dependence
of electron density changes from r \Gamma8 to r \Gamma4 at around 2 R fi . All these
suggest that the physics of the plasma transport and wave dissipation
diverges from classical Coulomb theory at heights beyond 1.5 R fi .
Acknowledgements
We would like to thank Prof. Ester Antonucci for her valuable participa­
tion in the joint UVCS and SUMER observation campaign. We would
also like to thank the anonymous referee for his comments. Research at
Armagh Observatory is grant­aided by the Dept. of Education for N.
Ireland while partial support for software and hardware is provided by
the STARLINK Project which is funded by the UK PPARC. This work
was partly supported by PPARC grant GR/K43315. SOHO is a project
of international cooperation between ESA and NASA. We would like
to thank the SUMER team at Goddard Space Flight Center for their
help in obtaining the present data.
sumer—plume—v4.tex; 16/11/1999; 11:49; no v.; p.16

Polar Plumes and Inter­plume regions 17
References
Antonucci, E., et al. 1997, ASP Conf. Series 118, 273
Banerjee, D, Teriaca, L., Doyle, J.G., & Wilhelm, K., 1998, A&A 339, 208
Charbonneau, P., 1995, ApJS 101, 309
Cranmer, S. R., Kohl, J.L., Noci, G., Antonucci, E., Tondello, G., Huber, M. C.
E., Strachan, L., Panasyuk, A. V., Gardner, L. D., Romali, M., Fineschi, S.,
Dobrzycka, D., Raymond, J. C., Nicolosi, P., Siegmund, O. H. W., Spadaro,
D., Benna, C., Ciaravella, A., Giordano, S., Habbal, S. R.,Karovska, M., Li, X.,
Martin, R., Michels, J. G., Modigliani, A., Naletto, G., O'Neal, R. H., Pernechele,
C., Poletto, G., Smith, P. L., & Suleiman, R. M., 1999a, ApJ, 511 481
Cranmer, S. R., Field, G.B., & Kohl, J.L., 1999b, ApJ, 518, 937
David, C., Gabriel A.H., Bely­Dubau, F., Fludra, A., Lemaire, P. & Wilhelm, K.,
1998, A&A 336, L90
DeForest, C.E., Hoeksema, J.T., Gurman, J.B., Thompson, B.J., Plunkett, S.P.,
Howard, R., Harrison, R., & Hassler, D.M., 1997, Solar Phys. 175, 393
Dodero, M.A., Antonucci, E., Giordano, S., & Martin, R., 1998, Solar Phys. 183, 77
Doyle, J.G., , Banerjee, D. & Perez, M.E., 1998, Solar Phys. 181, 91
Doyle, J.G., Teriaca, L. & Banerjee, D., 1999, A&A 349, 956
Esser, R., Fineschi, S., Dobrzycka, D., Habbal, S. R., Edgar, R. J., Raymond, J. C.,
Kohl, J., & Guhathakurta, M., , 1999, ApJ 510, L63
Hassler, D.M., Wilhelm K., Lemaire, P., & Sch¨uhle, U., 1997, Solar Phys. 175, 375
Kohl, J.L., Noci, G., Antonucci, E., Tondello, G., Huber, Cranmer, S. R., Strachan,
L., Panasyuk, A. V., S., Gardner, L. D., Romoli, M., Fineschi, S., et al., 1998,
ApJ, 501 L127
Koutchmy, S. & Bocchialini, K., 1997, in Robotic exploration close to the Sun; AIP
Conf. Proc. 385, 137
Lemaire P., Wilhelm K., Curdt W., et al., 1997, Solar Phys., 170, 105
Mariska, J.T., 1992, The Solar Transition Region. Cambridge University press, Cam­
bridge
McIntosh, S. W., Diver D. A., Judge P. G., Charbonneau P., Ireland J., & Brown
J. C, 1998, A&A S 132, 145
Moses, D., Clette, F., Delaboudini'ere, J.­P., et al. 1997, Solar Phys. 175, 571
Noci, G., Kohl, J.L., Antonucci, E., et al., 1997, Adv Space Res. 20, 2219
Peter H., 1999, ApJ 516, 490
Peter H., & Judge P. G., 1999, ApJ 522, L77
Siegmund, O.H.W., Gummin, M.A., Stock, J.M., Marsh, D., Raffanti, T., Sasseen,
T., Tom, J., Welsh, B., Gaines, G.A., Jelinsky, P. and Hull, J., 1994, Proc SPIE
2280, 89
Teriaca, L., Banerjee, D. and Doyle, J.G., 1999, A&A 349, 636
Tu, C.­Y, March, E., Wilhelm, K. & Curdt, W., 1998, ApJ 503, 475
Warren, H.P., Mariska, J.T., Wilhelm, K., & Lemaire, P., 1997, ApJ 484, L91
Wilhelm, K., Curdt, W., Marsch, E., Sch¨uhle, U., Lemaire, P., Gabriel, A., Vial,
J.­C., Grewing, M., Huber, M.C.E., Jordan, S.D., Poland, A.I., Thomas, R.J.,
Kuhne, M., Timothy, J.G., Hassler, D.M. and Siegmund, O.H.W., 1997a, Solar
Phys. 162, 189
Wilhelm, K., Lemaire, P., Curdt, W., Sch¨uhle, U., Marsch, E., Poland, A.I., Jordan,
S.D., Thomas, R.J., Hassler, D.M., Huber, M.C.E., Vial, J.­C., Kuhne, M., Sieg­
mund, O.H.W., Gabriel, A., Timothy, J.G., Grewing, M., Feldman, U., Hollandt,
J. and Brekke, P., 1997b, Solar Phys. 170, 75
Wilhelm, K., Marsch, E., Dwivedi, B. N., Hassler, D. M., Lemaire, P., Gabriel, A.
& Huber, M. C. E., 1998, ApJ 500, 1023
sumer—plume—v4.tex; 16/11/1999; 11:49; no v.; p.17