A coronal mass ejection (CME) is a significant and most dramatic component of coronal evolution. In addition CMEs provide major disturbances to the `quiet' interplanetary medium and, of course, can have a severe influence on the geomagnetic environment. We have known about CMEs for almost 30 years and have excellent extended series of observations of such events using instrumentation on spacecraft such as Skylab, Solar Maximum Mission, P 78-1 and the Solar and Heliospheric Observatory (SOHO) (see reviews by Hundhausen (1997) and Howard et al. (1997)). Despite this we are still not able to truly identify and agree upon the onset circumstances and mechanisms of these events.
This comes about for several reasons. First, CMEs are detected using coronagraphs, which necessarily occult the source regions of the CMEs themselves. Secondly, the coronagraph observations are best tuned to events in the plane of the sky. This means that any disc observing devices used to study the CME source region are directed to regions near the limb, which will suffer from extreme foreshortening or even be partly or fully occulted by the solar disc itself. Thirdly, to understand the physical processes taking place in the source regions at the time of a CME onset one had better use instrumentation capable of providing diagnostic information on the source region plasmas over a broad range of temperatures, i.e. a coronal spectrometer working in extreme ultraviolet (EUV) or X-ray wavelengths.
In addition, the differences between the fundamental emission processes in X-ray/EUV and visible light make comparisons extremely difficult. Visible light is detected in the corona as a result of the Thomson scattering of photospheric light off free electrons. Thus, the brightness is dependent on density (not temperature) and decreases with angle from the plane of the sky. X-ray/EUV emission from the solar atmosphere is dominated by spectral emission lines from highly ionised ions in the corona and brightness is a function of density (approximately density squared) and temperature. The emission is isotropic. Comparisons of the outward expansion of plasmas covering a large range of temperatures and densities, in visible light and X-ray/EUV radiation requires careful analysis.
For all of these reasons a multi-wavelength study has been set up and run on a number of occasions, where the ultimate aim is to provide EUV spectra of the onset phase of CMEs in the low solar atmosphere. We report here on one event observed during this campaign.
There have been many attempts to relate coronal features observed by disc-viewing instrumentation with outward propagating CME structures seen in white light. For example, Rust & Hildner (1976) and Hudson et al. (1996) identified outward expanding X-ray features from Skylab and Yohkoh, respectively, which they related to overlying CME structures. However, there is no clear one-to-one association between observed ascending X-ray features and white-light CME features and it remains to be shown that we can truly identify the CME plasma source prior to and during the onset of the CME eruption.
The principal feature of interest in this report is the dimming observed in coronal emission lines in association with CME events. Such dimming in association with a CME has been reported by Rust & Hildner (1976), and more recently by Sterling & Hudson (1997), Harrison (1997b), Gopalswamy & Hanaoka (1998) and Zarro et al. (1999).
Rust and Hildner described a filament eruption, with no associated flare. The X-ray depletion was estimated to signal the `loss' of over 1.3 x 1012 kg compared to the calculated mass of the overlying CME at 2 x 1012 kg. This mass estimate, combined with the position, size and speed of ascent, suggested that the X-ray feature was indeed the early ascent of the CME. The Sterling and Hudson, and Zarro et al. reports both examine the flare of April 7, 1997, using Yohkoh and SOHO observations, respectively. Pockets of dimming in coronal EUV emission were detected in the vicinity of the flare, deemed to be the source region of a CME seen as a `halo' (Earth-directed) event. Sterling and Hudson estimate that a `loss' of some 1011 kg was associated with the dimming, which was observed on the disc at two ends of an `S' shaped active region.
Hudson & Webb (1997) have reviewed the dimming signatures in X-rays detected using the Yohkoh spacecraft, associated with mass ejection events. They point out that coronal depletions were first described using visible light data from the High Altitude Observatory's K-coronameter (Hansen et al. 1974). Hudson and Webb describe several data-sets where there is evidence for the outward expansion of X-ray emitting material and associated dimming. Most involve flare activity. They present a basic classification scheme for X-ray dimming signatures which includes (i) dimming above a long-duration flare event, (ii) cloud ejections (an expanding X-ray cloud adjacent to a flare), (iii) streamer `blow out' events, and (iv) transient coronal holes (dimming of the corona near an X-ray arcade).
Most authors agree that the dimming signatures detected in X-rays and EUV may be critical for detecting the onset of CMEs and for investigating the physics of the onset process itself. In the opinion of the current authors, there is much confusion, for a number of reasons, and this makes investigations of the dimming process rather urgent. Of most concern are the following points:
1. Almost all of the observations of dimming which have been reported make use of wide-band imagers, e.g. SXT on Yohkoh, the X-ray spectrographic telescope on Skylab and EIT on SOHO. The plasma diagnostic information is necessarily limited and a spectroscopic campaign is urgently required.
2. Many reports (e.g. Sterling & Hudson, 1997, and Zarro et al., 1999) identify patches of dimming on scales much smaller than the associated active regions. However, the average CME extends over 45 heliographic degrees and the average active region may be one tenth of this. How can we relate such dramatically different spatial scales? CME spans are not only frequently larger than the associated active regions, they do not change with time, i.e. in most cases they will not project back to a region as small as an active region. If dimming is associated with the removal of mass in a CME, it would surely extend over a large area of the corona - commonly larger than an active region. How relevant are the sub-active region dimming patches to this?
3. Flares are highly dynamic events signaling a major disruption to the low corona in a localised area. One might expect to identify dimming in associated sites even if there was no CME association. Changes in temperature and the transport of plasma may well result in dimming in any case at certain locations in specific wavelengths. It would be better to examine dimming off the limb and away from the closed loop strutures of an associated active region when attempting to identify X-ray/EUV dimming in association with a CME.
4. Solar physics frequently suffers from a desire to classify events, i.e. to put events into boxes. This is often done without real physical justification. A good example of this is the so-called long duration event (LDE) (see Sect. 6 of Harrison, 1995). It remains to be seen whether the Hudson & Webb (1997) dimming classification leads to a better physical insight, but it is difficult to see how it can be useful when (i) it emphasises the flare (LDE) association so much, (ii) one could easily imagine many events satisfying all classifications, and (iii) the observation of dimming events is clearly in an early stage and is mainly defined by wide-band instruments.
It is the opinion of the authors that one needs to stand back, ignore any desire to classify events and make spectroscopic observations of dimming events off the limb to identify the physical processes at work. This is the object of the campaign being described in this work.
The principal tools of the campaign are the Coronal Diagnostic Spectrometer (CDS; Harrison et al., 1995) and the Large Angle Spectroscopic Coronagraph (LASCO; Brueckner et al., 1995) both aboard SOHO. The CDS instrument is an EUV spectrometer operating in the region 150-800 Å. Thus it can provide information on a large range of spectral emission lines from many trace ions in the solar atmosphere whose characteristic temperatures range from 20,000 K to 2,000,000 K and, for plasmas in this range, we can derive information on plasma density, temperature, flow speeds, abundances, as well as topology and evolution. LASCO is a three-component coronagraph system capable of detecting and tracking CMEs from 1.1 to 30 solar radii.
The operation of the CDS instrument in this observing campaign is rather complex, so the details are described in the next section. This is followed by an overview of the observations of 16 July 1997 and an analysis of the EUV data. We show that the CME of 16 July is associated with coronal dimming, activated cool features and some 2 million K `hot spots'. The results and the impacts of these on our understanding of CME onsets are then discussed.
© European Southern Observatory (ESO) 2000
Online publication: June 20, 2000