Ionospheric Electromagnetic Energy Input
Here we examine the electromagnetic energy input into the ionosphere by assessing the Poynting vector associated with perturbations along the satellite world-line, calculated using\(S=\ \frac{1}{u_{0}}E\ \times B\) where \(u_{0}\) is the magnetic constant, and E and B denote the electric and magnetic field vectors of the perturbation fields, respectively. By applying band-pass filters it is possible to remove the influence of large-scale variations of the Earth’s main field as measured along the trajectory of the moving satellite, and to focus on the Poynting flux contributions arising from perturbations at various scales of interest. With a single satellite it is impossible to uniquely separate the impacts arising from spatial and temporal variations along the satellite world-line. However, as shown for example by [16][17], analysis of the wave impedance as a function of frequency in the Swarm frame provides strong evidence for the importance of Alfvén waves in MIC.
Figure 1 shows the statistical Poynting flux over two separate one month-long time periods, one in the northern near-summer solstice conditions (1-31 July, 2016; panels (a) and (b)) and the other in the northern near-winter solstice conditions (15 November-15 December, 2016; panels (c) and (d)), for both the dayside (left column) and the nightside (right column) as determined by magnetic local time (MLT). These two intervals were chosen to reflect periods where the Swarm A orbits were in similar noon-midnight local time orientations. In this Figure, the error bars show the median spanned by the upper and lower quartiles in the statistics, with the scale dependence of the Poynting flux as a function of frequency derived by the application of a time-domain Savitzky-Golay low-pass filter of varying width along the x-axis (see Methods for details). It can be seen that on the dayside during near-summer solstice, there is a clear statistical preference for more electromagnetic energy to be driven into the northern hemisphere than the southern hemisphere at Swarm altitudes. On the dayside during near-winter solstice, the preferential direction of the energy transfer does reverse such that there is more Poynting flux directed into the southern hemisphere. However, and very significantly, the asymmetry in the interhemispheric energy transfer is much smaller than in the near-summer solstice period. As a result, there is a clear preference for more energy transfer into the north. Indeed, if the results from these two months approximating the near-summer and near-winter solstice periods are summed, the implied summer-winter seasonally-averaged Poynting flux will have a clear net northern preference.
On the nightside (panels (b) and (d)), the northern preference for electromagnetic energy transfer is even more stark. During the near-winter solstice on the nightside (panel (d)), there is a reduction in the northern preference, but remarkably the direction of the electromagnetic energy transfer does not reverse as compared to the near-summer solstice such that the median direction of energy transfer remains slightly northwards even in the near-winter solstice at night.
In all cases the error bars plotted in this Figure, and which refer to the 25% and 75% quartiles in Poynting flux, appear be a significant fraction of the median. However, we emphasise that this feature should not be interpreted as a low statistical significance of our result demonstrating northern preference for electromagnetic energy transfer seen at Swarm. Instead, the large range of average Poynting flux magnitudes represented by the quartiles simply reflects the expected large variability in the magnitudes of energy flux from hour to hour and day to day in response to non-steady solar wind forcing. Supplementary Material Figures 1 and 4, for the northern near-winter and near-summer solstice periods, respectively, show this effect clearly. It can be seen that during more intense geomagnetic activity, the magnitude of the Poynting flux increases in both hemispheres. This can be seen particularly during conjugate observations from adjacent northern and southern hemisphere passes, where the Poynting flux is seen to increase and decrease in tandem in both hemispheres in response to varying levels of driving. In particular for the near-summer solstice period it can be seen that the northern Poynting flux dominates over the conjugate southern hemisphere counterpart, both on the dayside and on the nightside, in the time domain across the whole interval despite it spanning a wide range of intensities of solar wind driving conditions. Therefore the northern preference for electromagnetic energy input persists in the time domain from event to event, and not just when combined statistically as in Figure 1.
Figure 1 further shows that in general the electromagnetic Poynting flux observed at Swarm appears to be smaller on the nightside than on the dayside. This is evident both in the near-summer and near-winter solstice periods and appears to be a general characteristic of the magnitude of the observed electromagnetic energy input arising from electromagnetic fluctuations at this altitude. A likely explanation for this is that a significant fraction of the incoming electromagnetic energy is converted to the kinetic energy of downgoing auroral electrons as a result of coupling at higher altitudes above Swarm in the nightside auroral acceleration region (AAR) located around 4000-12000 km in altitude [18]. This inference is consistent with the concept whereby the ionospheric feedback instability [19] can produce discrete arcs which convert incoming electromagnetic energy into field-aligned electron acceleration in the AAR. This feedback process happens preferentially at night where the background conductivity is low, and where in the absence of dayside EUV illumination strong conductivity gradients can be formed [20]. In such a paradigm, the reduction in nightside Poynting flux observed at Swarm, located below the AAR, is explained as a result of significant energy removed in association with discrete arc auroral electron acceleration above.
Interestingly, in the data shown in Figure 1, the interhemispheric energy fraction appears to be independent of scale. This suggests that the processes responsible for the observed asymmetry are most likely self-similarly active at and/or self-similarly impact all transverse scales considered.