The key scientific goals of SMILE mission are to investigate the dynamic response of the Earth's magnetosphere to the solar wind impact via simultaneous in situ solar wind/magnetosheath plasma and magnetic field measurements, X-Ray images of the magnetosheath and magnetic cusps, and UV images of global auroral distributions. Imaging of the magnetosphere and auroras for more than 40 hours continuously.

SMILE will answer the following science questions:

1) What are the fundamental modes of the dayside solar wind/magnetosphere interaction?

2) What defines the substorm cycle?

3) How do CME-driven storms arise and what is their relationship to substorms?

What are the fundamental modes of the dayside solar wind/magnetosphere interaction?

Dayside reconnection causes plasma to flow anti-sunward through the magnetopause boundaries, the cusps, and over the polar caps. On occasions reconnection can persist for long times, both for northward and southward IMF orientations (e.g. Frey et al. 2003; Phan et al. 2004). However, reconnection can also be bursty and time dependent, generating significant structure and step motion of the magnetopause (Figure 1). For example, patchy reconnection (Russell and Elphic 1978), bursty (i.e., time dependent) reconnection from a single X-line (Scholer 1988; Southwood et al. 1988), and multiple X-line reconnection (Lee and Fu 1985; Raeder 2006; Omidi and Sibeck 2007), may produce so-called flux transfer events (FTEs) which, at the most basic level, may be thought of as time dependent structures propagating along the magnetopause.

Figure 1: Location of the subsolar magnetopause as a function of time for slow reconnection (dotted line), bursty reconnection (red line), normal rate reconnection (grey line) and fast reconnection (solid line).

Reconnection is therefore thought to cause the shape of the magnetopause to become blunter. By contrast, variations in the solar wind dynamic pressure should cause self-similar changes in magnetospheric dimensions. Thus by measuring the curvature, size and absolute location of the magnetopause and the location (latitudinal position), size, and shape of the cusps, it is possible to distinguish the differing effects of pressure changes and magnetic reconnection on the global magnetospheric system. This would distinguish on a global level the nature of the solar wind - magnetosphere interaction, the dominant driving mechanisms and modes of interaction.

Figure 2：Cartoon showing the progression of the Dungey cycle. Under southward IMF conditions, dayside reconnection (panel A) opens magnetic flux (panel B) which convects over the poles and is stored as magnetic energy in the magnetotail lobes (panel C). This stored energy accumulates until an explosively release (panel D) returns closed flux to Earth in conjunction with dramatic auroral displays at high latitudes (panel E).  Substorms may result from changes in the external driving of the magnetosphere and/or internal magnetotail instabilities. (From Eastwood et al. 2015)

What defines the substorm cycle?

We know that southward IMF is required to increase the energy density of the magnetotail lobes, and the more prolonged the interval of southward IMF, the more energy is stored, but the precise nature of the energy loading and the role it plays in the subsequent onset of geomagnetic activity is very controversial. For example, one very fundamental question is whether each substorm requires its own interval of loading (growth phase), or whether multiple substorms can occur in response to a single growth phase.

The polar cap is an area of magnetic field lines that are open to the solar wind and is readily identified by the auroral oval which bounds it (Figure 3). Auroral oval observations provide information about the ionospheric footpoints of magnetopause processes. Specifically, the expanding-contracting polar cap paradigm utilises basic properties of the auroral oval to provide direct measurements of the state of the magnetosphere by measuring the size of the polar cap (e.g. Milan et al. 2009). The area of open flux within the polar cap changes directly in response to the amount of open flux in the magnetotail lobes, and the very dynamic changes that occur in this region are in response to different solar wind conditions.

Figure 3: (a and b) Determination of the size of the polar cap from the radius of the auroral oval; (c, d, and e) time-series of the radius of the auroral oval, the Sym-H index measure of ring current intensity, and a proxy for the dayside reconnection rate derived from upstream solar wind conditions.  (Milan, 2009). These measurements have necessarily been non-continuous in the past due to the orbits of previous auroral imaging missions (data gaps in panel c), hindering progress in understanding solar wind - magnetosphere coupling. Short time-scale variations in polar cap size correspond to substorms, but large discontinuities exist over some data gaps indicating that the storm behaviour is only partially captured.

A recent development is the awareness of a low-intensity auroral feature called auroral beads that develop in pre-breakup auroral arcs and eventually produce the initial brightening and substorm expansion onset (Henderson, 1994). This feature was not recognized earlier due to its low intensity and is shown in Figure 4. The auroral beads have specific wavelengths and corresponding exponential growth in the auroral intensity that are different from case to case, apparently dependent on the state of the magnetosphere just prior to substorm expansion onset. The characteristics of auroral beads revealed recently impose another set of rather severe observational constraints that discriminate among several potential substorm onset processes under consideration. Two potential plasma instabilities that may account for these characteristics are the ballooning instability and the cross-field current instability. The latter was recently examined and was found to account for the observed auroral bead characteristics, as illustrated in Figure 4. More recently, however, Kalmoni et al. (2018) showed that a third mechanism, kinetic Alfve？n waves, could explain the temporal and spatial scales of auroral beads. The coordinated global imaging from SMILE and ground-based auroral observations around substorm expansion onset would be ideal to test these proposed plasma instabilities further.

Figure 4. A schematic diagram to show (top) the conjugacy of auroral beads by ground-based observations from both Northern and Southern hemispheres, (middle) the wavelengths in the magnetospheric equatorial plane corresponding to auroral beads, and (bottom) the current filamentation and electron acceleration arising from the excitation of the cross-field current instability

Disentangling these different modes of behaviour follows on from the first question. Once a substorm is triggered, what controls its subsequent evolution? To what extent is it sensitive to changes in the solar wind conditions, and how does this sensitivity depend on the internal state of the magnetosphere (e.g. substorm phase, amount of remaining stored energy, etc.)?

How do CME-driven storms arise and what is their relationship to substorms?

While intervals of southward IMF occur naturally in the solar wind, and so substorms occur on a daily basis (Borovsky et al. 1993), strong driving causing geomagnetic storms tends to occur in response to coherent solar wind structures, particularly Coronal Mass Ejections (CMEs) (Gonzalez et al. 1999).

The degree to which solar wind plasma, momentum and energy enter the magnetosphere is characterized by so-called solar wind coupling functions (Gonzalez 1990; Finch and Lockwood 2007). Physically, magnetic reconnection at the dayside magnetopause is enhanced if there is a strong interplanetary magnetic field component opposite to the dayside magnetospheric magnetic field, supplemented by fast solar wind, for an extended period of time.

CMEs are transient eruptions of material from the Sun’s corona into space (Forbes 2000). CMEs propagate at super-magnetosonic speeds relative to the ambient solar wind, and play a particularly important role in the dynamics of the Earth's magnetic field because they can contain long intervals of southward IMF (e.g. Gonzales et al. 1999). In general, the largest geomagnetic disturbances are associated with CMEs, with the level of activity being directly related to the flow speed, the field strength and the southward component of the magnetic field (Richardson et al. 2001).

Figure 5：Coronal mass ejection (CME), shown as an orange arch travelling toward Earth. Left insert shows the CME observed by SOHO spacecraft and right inserts shows the aurora observed from space with Polar spacecraft (top) and from ground (bottom) (credit NASA).

Sometimes CMEs don’t have the expected effects associated with a geomagnetic storm. When the interplanetary magnetic field is northward the energy transmitted to the magnetosphere is more limited. However, when solar filaments are contained in CMEs, there can be some effects similar to superstorms such as the superfountain in the equatorial ionosphere, magnetotail stretching and strong joule heating in the polar ionosphere (Kozyra et al., 2014). Furthermore, Turc et al. (2014) showed that the Earth’s bow shock can, under certain conditions, modify the interplanetary magnetic field direction contained in CMEs which then do not have the predicted effect on the magnetosphere.

Understanding the global CME/magnetosphere interaction is crucial to understanding precisely how the structure of the CME is responsible for the different phases of geomagnetic storms. On a practical level, storms driven by CMEs have potentially severe space weather consequences and represent a significant threat to infrastructure resilience worldwide.

Very basic questions still remain. Is the duration and magnitude of solar wind driving the sole arbiter of whether a storm will occur? What is the relationship between the storm and substorm? Are storms always a separate phenomenon, or can they be considered as being composed of multiple substorms?

Finally, although the question of how a storm starts has been central to the scientific studies of the magnetosphere for as long as measurements have been available, the question of duration is growing in importance, driven by the needs of the end-user in the space weather context (i.e. confidence in issuing ‘all clear’). Does a storm end because it has exhausted the reservoir of stored magnetic energy in the magnetotail? Or does a storm stop because the solar wind driving conditions have changed? If both possibilities are observed to occur, which is the more important? And once the solar wind driving is removed, how rapidly does the magnetosphere recover? Is it more likely that the solar wind conditions will change, or is the stored magnetotail lobe energy depleted so rapidly that the changing solar wind plays only a minor role?