Aims. There have been many studies of extreme-ultraviolet (EUV) dimming in association with coronal mass ejection (CME) onsets. However, there has never been a thorough statistical study of this association, covering appropriate temperature ranges. Thus, we make use of a large campaign database utilising the Coronal Diagnostic Spectrometer (CDS) and the Large Angle and Spectrometric COronagraph (LASCO) both on the SOlar and Heliospheric Observatory (SOHO) to associate dimming events detected at 1 and 2 million K with CME activity. The aim is to confirm whether the dimming-CME association is real or not. This in turn will confirm whether special attention should be paid to the EUV dimming in the pre-eruption and eruption periods to study the CME onset process itself.
Methods. The CDS CME onset campaign data for Mg IX and FE XVI observations on the solar limb are used to compare to LASCO event lists over a period from 1998 to 2005. Dimming events are identified and the physical extent explored, whilst comparing the events to overlying CME activity.
Results. For the identified dimming regions we have shown strong associations with CME onsets, with up to 55% of the dimming events being associated with CME activity. This is compared to the random case where up to 47% of the dimming regions are expected to be associated with CMEs. We have also shown that up to 84% of CMEs associated with our data can be tracked back to dimming regions. This compares to a random case of up to 58%.
Conclusions. These results confirm the CME-EUV dimming association, using a statistical analysis for the first time. We discuss the repercussions for the study of CME onsets, i.e. analysis of the dimming regions and the periods up to such dimming may be key to understanding the pre-CME onset plasma processes. The results stress that one emission line may not be sufficient for associating dimming regions with CMEs.