5 Summary and conclusions

CDS spectroscopic observations made on the solar disk have been used to characterise polar plumes, comparing them with earlier observations of a low-latitude plume, seen in the Elephant's Trunk coronal hole near disk-centre (Del Zanna & Bromage 1999b). These plumes are all diffuse, quasi-stable structures, rather than the category of plume which has a bright, hot footpoint. The latter may in fact be an earlier stage in the evolution of the plume. Here, we have presented the results from two plumes, typical of the many such diffuse structures to be seen in the polar coronal holes around solar minimum. They exhibit the same characteristic signature as the low-latitude plume, namely, enhancement of the cool, upper transition region lines such as Mg VII, no emission in the hotter coronal lines (formed above $1.2\times 10^6$ K) and an order-of-magnitude increase in ratios of high-FIP to low-FIP lines (e.g. Mg VII/Ne VII).

We have presented the first GIS observations of the base of a plume, and used them to confirm the results obtained from NIS observations. In particular, we have used them to better define the DEM and the elemental abundances. We have shown that the DEM technique, combined with the large number of observed lines and a good instrument calibration, is extremely useful for obtaining information on relative element abundances and on the temperature distribution along the line of sight. We have shown that other methods, widely used, can lead to inaccurate results. However, the results of the DEM analysis depend critically on the ionization fractions adopted. We have pointed out that the currently available ionization equilibrium calculations for a significant number of ions lead to large inconsistencies. Unfortunately, it has been common practice to use lines from ions with anomalous behaviour for diagnostic studies of solar and stellar atmospheres. This has had serious consequences in the derivation of densities and element abundances from stellar EUV spectra as shown by Del Zanna et al. (2002). Any previous results on the solar atmosphere based on the use of these ions should therefore be taken with caution. More refined emission measure and element abundance studies will only be possible when the problems in the ionization equilibria have been solved.

Both line ratio techniques and DEM analysis have shown plume plasma to be close to isothermal, with a temperature $T \simeq 7{-}8 \times 10^5$ K, a little cooler than the surrounding coronal hole region. This was found using a number of different line-ratios and was also the temperature at which a sharp peak in emission measure can be seen in the DEM curve. The less isothermal emission from the surrounding coronal hole region can be seen as a broader peak in the DEM, close to, or just below, $1\times 10^6$ K.

For any theoretical modelling, it would be important to study in detail any temperature variations across and along the length of the plumes. The observations presented here indicate a small increase in temperature with height along the plumes and no variation across them. A more detailed study will be the subject of a future paper. Here we only mention that Del Zanna (1999) has shown that plumes have shallower radial temperature gradients compared to the inter-plume regions, and that these temperature differences tend to disappear with distance, as all the coronal hole plasma becomes almost isothermal.

Accurate density measurements are difficult to obtain, because of the weakness of the density-sensitive lines. At lower-coronal heights ( $T \simeq 8-9 \times 10^5$ K; e.g., Mg VIII, Si IX) they have rarely shown significant differences between plume and inter-plume regions ( $N\mbox{$\rm _{e}$ } \simeq 2 \times 10^{8}$ cm^-3). However, due to low signal, the uncertainties are large and do not exclude the possibility of slightly higher density in plumes. At lower temperatures ( $T \sim 7 \times 10^5$ K), where plumes have their maximum emission measure and visibility, the density (Mg VII) is $N\mbox{$\rm _{e}$ } \simeq 1.2 \times 10^{9}$ (cm^-3), higher than in the coronal hole regions, but similar to quiet sun values. At even lower temperatures ( $T \sim 2 \times 10^5$ K), in the lower transition region (O IV), plumes are not visible, but O IV emission at the base of the plume indicates densities which are the same as in the nearby coronal hole network regions ( $N\mbox{$\rm _{e}$ } \simeq 5 \times 10^{9}$ cm^-3).

It has been shown that the characteristic line intensities in the plume are mainly a result of their quasi-isothermal temperature structure. In particular, the large enhancement of some intensity ratios (such as Mg VII/Ne VII) seen at the base of the plumes can largely be explained by this temperature effect, rather than implying large changes in abundance.

We have not found any strong evidence for an FIP effect in plumes, contrary to earlier reports. We have shown that the Skylab plume observations can similarly be explained by the plume's temperature characteristics. However, in the plumes we observed, we have found a consistent small reduction in Ne abundance (relative to photospheric values). The result that plumes do not after all exhibit any significant FIP effect means that the previous deduction (see the introduction), that plumes could not be the source of the fast solar wind, is no longer valid. However, our result does not prove that they are the source of the fast wind. Indeed, as described in the introduction, other recent studies suggest that the source region is more likely to be between the plumes.

We have used the plume GIS observations to show that the EIT filter response is not isothermal. In particular, we have shown that, in the case of plumes, the low-temperature lines provide the dominant contribution to the EIT filter images. As a result, these filters cannot be reliably used for temperature estimates without an in-depth analysis. We note that similar arguments apply to other broad-band filter ratio techniques such as those based on Transition Region And Coronal Explorer (TRACE) data, that have been widely used in the last few years. This is because TRACE filters have similar responses to the EIT ones. All the above shows the importance of complementing any intensity measurements with spectroscopic observations in order to determine the physical parameters of various solar coronal features. Care must also be taken that the analysis technique adopted is appropriate for the conditions of the observation.

Acknowledgements

G. Del Zanna and H.E. Mason acknowledge support from PPARC. The authors thank Jeff Newmark for providing EIT images and effective areas, and the CDS team for support in the observations. GDZ was also supported by a University of Central Lancashire studentship for part of this work. SOHO is a project of international collaboration between ESA and NASA.