The old concept that solar flares and, more recently, isolated filament eruptions cause major transient interplanetary disturbances and geomagnetic storms is being challenged. Many researchers embrace a new paradigm that the CME, a subset of which are associated with both flares and filament eruptions, is the crucial link between solar activity and transient disturbances at Earth. Much of this misconception arose because CMEs are eruptions of coronal, not lower atmospheric material. To earlier researchers the corona was mostly invisible, whereas flare effects in the lower solar atmosphere were easily viewed in the optical regime. The development of spaceborne imaging instruments allowed us to study the corona and eventually correct our understanding of solar-terrestrial linkages. The concept that CMEs are the key geoeffective solar phenomenon has important implications for prioritizing the development of instruments and techniques for the observation of signatures at the sun and in the vicinity of the earth used to predict transient IP disturbances and geomagnetic storms.
CMEs arise in large-scale structures lying near the base of the solar corona. Most of the ejected material originally consisted of hot, highly ionized gas imbedded in closed loop structures. Most CMEs are associated with coronal streamers, many of which are in the streamer belt that is the base of the HCS. Thus, CMEs may be an integral part of dynamical phenomena occurring in the HCS. Two important implications of this are that the transient disturbances related to CMEs will tend to occur more frequently near sector boundary crossings, and that the stronger recurrent storms, usually thought to be caused by high speed streams between sector crossings, may actually be associated with compression of CME flows in stream interaction regions near the boundaries [ Crooker and Cliver, 1994]. Therefore, determining the time of sector boundary crossings at the earth, based on the location and degree of activity in the streamer belt at the sun, should improve storm forecasts.
The largest storms are caused by fast CMEs and strong shocks which often have associated energetic flares at the sun, but most storms are of moderate to small size. The disturbances that cause these storms have few clear observables at the sun, partly because we don't directly observe against the disk the coronal plasma and magnetic fields participating in the CME. In the optical regime disappearing filaments should be the best predictor of CMEs, but they are not well observed or routinely reported. In general, predictions should be based not on the occurrence of a large flare, but whether the observed activity is indicative of a CME.
Therefore, high priority should be given to the development of instrumentation to detect the onset of CMEs at the sun and their propagation through the corona and the heliosphere. Some older techniques have proven vital to our present understanding of coronal and IP mass ejecta, and should be considered in future planning. These include X-ray imagers, white light coronagraphs and metric radio telescopes for observations near the sun. Near-Earth coronagraphs, of course, best observe material above the solar limb and not CMEs directed at the earth. Observations of ejecta in the IP medium and the vicinity of the earth have been made by heliospheric white light imagers, detectors observing IP radio scintillation, kilometric radio telescopes and spacecraft instruments measuring the in-situ solar wind.
In addition to careful application of these known technologies,
we need to develop new methods to study CMEs and forecast their
arrival at Earth. Some promising missions involve directly
imaging dense plasma within the inner heliosphere. These include
white light imagers placed in Earth orbit, one or more
coronagraphs placed at the Lagrangian points at 60
or at
about 90
leading or trailing the earth in its orbit, and
instruments designed to image IP shock waves. These methods are
intended to provide 1-3 days advance warning of a disturbance
headed toward the earth.
Acknowledgments. I thank E. Cliver of AF Phillips Laboratory/GPS for useful discussions, and G. Siscoe and the referees for helpful comments on the manuscript. The work of DFW was supported at Boston College by the Geophysics Directorate of Phillips Lab under contract AF19628--90--K--0006.