6. Summary and discussion
For the event under study here, we witness the activation, but non-eruption of a large prominence adjacent to a compact CME which appears to lie over a small active region.
The most significant EUV feature under the span of the ascending CME is the dimming of 1,000,000 K plasma. Other temperatures do not show a similar feature, suggesting that the dimming is due to the decrease in density of plasmas at 1,000,000 K. The dimming is continuous and gradual. There is no clear onset - the depletion is underway from the start of the observations - and there is no clear `event' which may coincide with a CME onset, i.e. the drama of the eruption higher in the corona is not duplicated in the low corona, where we see a subtle, gradual effect. The dimming certainly starts before the CME crosses above the occulting disc and continues well after the CME onset. Indeed, the dimming appears to start before the projected CME onset.
The dimming is consistent with the loss of significant mass, which could account for at least 70% of the mass of the CME. This stresses that the dimming events are most likely the source regions of at least a large fraction of associated CME masses. This being the case, one would suspect that several features of the dimming will ultimately lead to a better understanding of the CME onset process. CME onset models must explain the gradual loss of mass in the low corona, at particular temperatures. There appears to be no sudden onset to the dimming and it may be ongoing before the projected onset of the CME seen in the coronagraph data.
At the northern footpoint of the CME/dimming, which is the southern extreme of the activated prominence, we find a compact 2,000,000 K blob, which is reminiscent of the compact X-ray emissions associated with eruptive events reported by Harrison et al. (1988). Thus, we have two large structures adjacent to one another, both of which become active, simultaneously (the CME and the activated prominence) and, at a site between them, we witness a bright `hot spot'. This scenario suggests that we are witnessing the interaction of two large-scale magnetic structures.
A second `hotspot' is located under the southern footpoint of the CME later in the sequence. Is it merely coincidence that a patch on the disc at the southern extreme of the CME/EUV dimming region brightens at these temperatures late in the event? It is certainly suggestive of localised heating resulting from the adjacent eruptive activity.
The initiation of a CME can be gas driven or magnetically driven. In the former, we would be looking to processes such as magnetic buoyancy or plasma pressure (due to some kind of density/temperature pulse). In the magnetic case, we might consider an ideal instability (non resistive conditions) or non equilibrium, or magnetic reconnection (resistive conditions).
Let us consider this event alone, with no reference to former events and analyses. A plasma pressure driven CME would require a very significant density/temperature increase at the base of the corona, and this we just do not see. There is no flare associated with this event, for example. In the case of magnetic buoyancy, it would be very unlikely that a structure would ascend, leaving a region of depleted density (the dimming); mass has to support the buoyant structure and this is not consistent with the dimming. Thus, we believe that this event points towards a magnetic driver. We note that the activation of a large adjacent structure (the prominence) and the hotspot are possible indicators of the interaction of large loop systems. This fact, combined with the lack of observation of any transient low-lying event at any of the observed temperatures suggests that either magnetic reconnection is occurring relatively high in the corona in a similar manner to that described in the magnetic breakout picture of Antiochus et al. (1999) or that there is an instability or non-equilibrium resulting in an ascending structure and subsequent interaction with the adjacent structure (activated prominence) as the event progresses.
We believe that the fact that the dimming is gradual is important. There is no clear onset of any kind in the dimming. One can imagine a scenario where the evolution of the magnetic system is such that we witness slowly ascending loops, for example, in response to magnetic shear of an arcade. The loop tops may be high in the corona, and the low coronal signature is simply registered in the EUV as a change in density. This can be taking place well before the CME actually takes off. At some point either the magnetic instability or non-equilibrium is reached, or reconnection has taken place as the magnetic structure interacts with an adjacent large structure. No front will pass through the principal dimming region because it is well below the loop fronts. The dramatic ascending loop front which leads the CME will be associated, below, with only further dimming. Of course, if a prominence was to erupt behind the CME front, this would be seen crossing the low corona as cool material.
Why should the density decrease be associated with a specific temperature? The answer to this might be simply the fact that the loops which are ejected contain principally plasma at that temperature and the line of sight effect is to see a fall in intensity at, say, 1 million K, while other temperatures show little decay in intensity. Alternatively this may be telling us something more fundamental about the ejection process. For example, there may be a resonant acceleration process, such as `surfatron' acceleration as a wave passes throught the corona, whereby plasma at a specific temperature is accelerated.
So, the dimming appears to be an extremely important signature of the CME onset process for the discrimination between CME models. This has important implications for coronal evolution studies and studies of mass loss processes on the Sun and all stars, but it may also be a valuable tool for the prediction of CME activity. On the limb, we detected EUV dimming prior to the detection of the CME in the coronagraph. A good test of our predictive capabilities would be to scan EUV images of the limb and predict the detection of subsequent CMEs in the corona above. Perhaps even more important would be to translate this to the prediction of CMEs on the solar disc, from EUV dimming. This would allow the prediction of Earth-directed CMEs which are not always readily detected using coronagraphs.
It is important to remember that we have described only one event here, in an effort to demonstrate the nature of the dimming process. This must be followed up with further spectroscopic analyses of dimming events in order to establish the typical characteristics and basic nature of these events.
© European Southern Observatory (ESO) 2000
Online publication: June 20, 2000