Along with the magnetic reconnection process, the solar wind and the interplanetary magnetic field (IMF) drive a large-scale plasma and magnetic flux circulation, convection, in the Earth’s magnetosphere-ionosphere system, the geospace. In a simplified picture, postulated first time already 60 years ago ( Dungey, 1961), during southward IMF, the southward IMF field lines reconnect with the northward closed geomagnetic field lines at low latitudes on the dayside magnetopause. The newly created open field lines are then dragged antisunward across the polar caps to nightside into the magnetotail by the solar wind. There, in the magnetotail, the magnetic field lines are reconnected again, and subsequently created closed field lines are convected earthward and via dusk or dawn back to the dayside where they become ready to reconnect with the IMF again. In the high-latitude ionosphere, this circulation is seen as two large-scale convection cells, in which plasma flows antisunward from noon to midnight over the polar cap and returns back to the dayside at lower latitudes roughly within the dusk and dawn auroral oval.

During purely northward IMF, the situation is more complicated. Convection is generally much weaker, and plasma and magnetic flux circulation occur preliminary in the magnetotail lobes. The northward directed IMF reconnects with the open magnetic field lines of the high-latitude tail lobes ( Dungey, 1963). The resulting new open field lines are draped over the dayside magnetopause until they are eventually dragged tailward by the solar wind ( Crooker, 1992). The corresponding ionospheric convection consists of 2 cells in the polar cap with sunward flow in the centre and antisunward flow just poleward of the dusk- and dawnside auroral oval ( Dungey, 1963; Maezawa, 1976). Most of the time however, the magnitude of the dusk-dawn IMF component (IMF B _y ) is larger than the northward directed north-south IMF component (IMF B _z ), that is, IMF |B _y | > IMF B _z ( Zhang et al., 2019). For such a dominant IMF B _y conditions, dayside magnetopause reconnection may occur even in a case of northward IMF B _z ( Nishida et al., 1998; Lee et al., 2010), or the two reconnection modes may co-occur ( Sandholt et al., 1998). Thus, an essential driver of magnetospheric convection is dayside magnetopause reconnection.

While it is well established that IMF B _z controls the strength of the convection, lesser is known on the role of dusk-dawn IMF B _y . However, we know that a presence of nonzero IMF B _y can cause plasma flow asymmetries in the high-latitude ionosphere, which manifest in that the two-cell convection pattern is distorted (e.g., Reiff and Burch, 1985; Heppner and Maynard, 1987). In the nightside, depending on the IMF B _y direction, one of the convection cells, either the evening or the morning cell, can extend significantly to the midnight sector (e.g., Ruohoniemi and Greenwald, 2005). Such deformations in convection are generally mirrored between the hemispheres along the noon-midnight meridian.

If asymmetry in ionospheric convection under the influence of nonzero IMF B _y exists, the key question is then how is the corresponding magnetospheric convection affected? Previous studies have shown that prevailing prolonged nonzero IMF B _y can affect the magnetotail by twisting it: The plasma sheet/cross-tail current sheet can get rotated around its axis (e.g., Kullen and Janhunen, 2004; Tsyganenko et al., 2015; Pitkänen et al., 2021a)) and an additional B _y component collinear to IMF B _y can be induced to the tail field to an extent that the closed tail field lines are bent or twisted from the meridian direction (e.g., Kaymaz et al., 1994; Tenfjord et al., 2015). These deformations in the magnetotail configuration, along with the observations of asymmetric ionospheric convection suggest that under nonzero IMF B _y , the magnetospheric plasma transport processes in the magnetotail plasma sheet could also be affected.

First observations-based indications of that nonzero IMF B _y conditions could affect convection also in the magnetotail plasma sheet were observed by Grocott et al. (2007) and Walsh et al. (2009). Based on simultaneous space-based (ESA’s Cluster satellites) and ground-based observations (SuperDARN radars), the authors reported events in which fast earthward ion flows perpendicular to the magnetic field in the near-Earth magnetotail plasma sheet were consistent with the hemispherically asymmetric fast convection in the nightside auroral oval. The perpendicular flows in the dusk-dawn direction in both the magnetotail and in the ionosphere were suggested to be a consequence of a (rapid) untwisting of magnetic field lines, following a reconnection in an IMF B _y -induced twisted magnetotail.

The observations by Grocott et al. (2007) and Walsh et al. (2009) motivated us to study more the IMF relationship with the magnetotail convection in more detail. In a statistical investigation of Cluster data, Pitkänen et al. (2013) discovered that the dusk-dawn component of the fast earthward perpendicular ion flows (>200 km/s) in the midnight near-Earth magnetotail (<20 R_E distance, where R_E is Earth’s radius) statistically correlate with the IMF B _y direction and tends to be opposite above and below the neutral sheet (B _x = 0), i.e., in the northern and southern plasma sheet. With an expanded dataset (Cluster + NASA’s THEMIS satellites, <30 R_E), Pitkänen et al. (2017) further reported that the mechanism that causes the interhemispheric flow asymmetry appears could work at all so-called IMF clock angles, where the IMF B _y direction is the critical deciding parameter.

Convection in terms of occurrence in the magnetotail plasma sheet is dominated by slow flows (<100 km/s, e.g., Chong et al., 2022). It is thus equally important to investigate how the slow flows are affected by IMF B _y . Pitkänen et al. (2018) addressed this by studying THEMIS measurements from a period of time, which showed clear signatures of an IMF B _y influence on the tail magnetic field. They found that both the earthward and tailward slow (<200 km/s) perpendicular ion flows were affected analogously as previously reported for earthward fast flows. By analyzing Cluster, THEMIS and ISAS’s/NASA’s Geotail satellite data, Pitkänen et al. (2019) statistically demonstrated the existence of interhemispheric asymmetry in average slow perpendicular flows. On the average, under clearly nonzero IMF B _y (IMF |B _y | > 3 nT), one magnetic hemisphere is dominated by a dusk-dawn flow component, which is oppositely directed compared to the other hemisphere. Under clearly positive IMF B _y conditions, the region of the earthward flows with a dawnward velocity component is extending to the midnight sector and expanding more and more to the premidnight sector with increasing tail distance in the northern plasma sheet (See Figure 1 demonstrating this (Figure 4 of Chong et al. (2022))). Similarly, the region of the earthward flows with a duskward velocity component is extending to the midnight sector, expanding more and more to the postmidnight sector with increasing tail distance in the southern plasma sheet. An analogous interhemispheric asymmetry but with an opposite sense is found for the flows under clearly negative IMF B _y conditions. Such an asymmetry is also observed for tailward directed ion flows in the near-Earth tail region (<32 R_E distance, Figure 1 (Figure 4 of Chong et al. (2022)). Comparison with the magnetic field indicated that the appearance of the dominating dusk-dawn flow component agrees with the appearance of tail By, which is interpreted to be induced by IMF By.

Research on this topic is progressing rapidly. Recently, Chong et al. (2022) investigated the slow perpendicular (to B) tail ion flows using a similar dataset as Pitkänen et al. (2019), which was expanded by the NASA’s MMS satellite measurements. Chong et al. (2022) focused on how the interhemispheric flow asymmetry under the influence of clearly nonzero IMF B _y depends on the distance measured from the neutral sheet. They used plasma ion beta (ratio of ion thermal pressure to magnetic pressure) as a proxy for the distance to the neutral sheet. Chong et al. (2022) found that the influence of IMF B _y on both the B _y component of the tail magnetic field and the perpendicular flow is more prominent in the midnight sector (compared to both the pre- and postmidnight sectors) and at distances far from the neutral sheet (compared to the distances close to the neutral sheet). The reason for these differences is not yet fully understood. The differences in the average flow patterns between the distances far from and close to the neutral sheet are demonstrated in Figure 1 (Figure 4 of Chong et al. (2022)).

Lane et al. (2022) studied fast earthward perpendicular flows by utilizing a vast dataset of Cluster, THEMIS and Geotail data. They focused on flows that had a velocity component toward the midnight meridian to distinguish them from the flankward diverging “symmetric” flows that are obtained when averaging the flow data without categorizing the magnetic hemispheres and the IMF B _y direction. In this approach, Lane et al. (2022) removed the contribution of such flows from their dataset that would formally be consistent with the untwisting hypothesis, but not necessary due to untwisting, something that might have affected the results by Pitkänen et al. (2013; 2017). Lane et al. (2022) found that ∼70% of the fast flow detections exhibit consistency with what would be expected according to the untwisting hypothesis (Agree flows) and ∼30% not (Disagree flows). They concluded this to indicate only a rather modest level of IMF B _y control. Lane et al. (2022) could infer that Agree (Disagree) flows tended to be accompanied by a localized perturbation to tail B _y in the same sign as (opposite to) the prevailing IMF B _y conditions, which temporarily enhances (overrides) the IMF B _y influence, see their superposed epoch analysis results in Figure 2 (Figure 3 of Lane et al. (2022)). Furthermore, Agree (Disagree) flows tended to be observed at larger (smaller) tail B _x , which suggest that they occur farther away from (closer to) the neutral sheet (B _x = 0). This is in accordance with the results by Chong et al. (2022) in which the IMF B _y influence on the slow flows was found to be more prominent farther away from the neutral sheet. The average slow “background” flows were found to be consistent with the untwisting hypothesis irrespective of whether the fast flow itself was Agree or Disagree, which is in accordance with the findings by Pitkänen et al. (2019) and Chong et al. (2022).

The oppositely directed dusk-dawn perpendicular velocity components in the flows in the northern and southern plasma sheet under the influence of nonzero IMF B _y indicate an existence of a reversal in the dusk-dawn velocity somewhere near or at the tail neutral sheet (B _x = 0). While implicitly present in one fast flow event discussed by Grocott et al. (2007) and noted in the other event analysed by Walsh et al. (2009), the actual velocity reversal has invoked only a very little attention. Walsh et al. (2009) discussed that the velocity reversal in their fast flow event could be related to the flows of the untwisting process, but could not draw definite conclusions. Pitkänen et al. (2015) reported another fast flow event with direct measurements of a dusk-dawn velocity reversal within the flow. The dusk-dawn velocity directions above and below the neutral sheet in the reversal as well as the tail magnetic field configuration and concurrent ionospheric convection were all consistent with those expected in the magnetotail which is influenced by IMF B _y .

Recently, Pitkänen et al. (2021b) investigated such earthward fast flow events measured by MMS, which were associated with clear dusk-dawn velocity reversals. All four analysed fast flow events were associated with signatures of the IMF B _y influence on the tail magnetic field. Three events were associated with dusk-dawn velocity reversals at the neutral sheet whereas one event was associated with reversals without any crossings of the neutral sheet. In those three events, the north-south component (E _z ) was the relevant convection electric field component and with the major contribution to both earthward and dusk-dawn perpendicular velocity components, the dusk-dawn components being consistent with velocities expected under the IMF B _y influence. The fourth flow event was a conventional fast flow with the dusk-dawn E _y and Sun-Earth E _x components contributing to X and Y perpendicular velocity components, respectively. It did not show clear consistency with what would be expected in nonzero IMF B _y conditions although the prevailing IMF B _y was strongly nonzero. These results suggest that when IMF B _y is influencing the magnetotail fast convection, then one can expect that the E _z electric field component will play a major role. This is supported by the electric field measurements in slow convection in a clearly twisted tail magnetic field configuration ( Pitkänen et al., 2018). The reason is that the magnetic field will be twisted or bent from the meridian direction and the convection electric field is by definition perpendicular to the magnetic field.

The relationship between the relevance of the different convection electric field components in fast flows and the IMF B _y influence has further been studied by Pitkänen et al. (2023). By investigating MMS measurements, Pitkänen et al. (2023) focused on the earthward perpendicular fast flows which fulfilled the frozen-in criterion. They found that the majority of the fast flow events in their dataset (52%) had E _z as the most relevant or dominating electric field component and only 26% of the events were conventional-type fast flows with E _y and E _x as the relevant components. Pitkänen et al. (2023) also found the IMF B _y influence on the fast flows to be more efficient as the relevance of E _z in the fast flows increases. This is consistent with the idea that the E _z convection electric field component should play a major role in a twisted magnetotail, as discussed above.

One related open question is in which time scale the tail responds to changes in the IMF. This has been addressed by using different approaches and methods like direct point measurements ( Motoba et al., 2011; Rong et al., 2015; Pitkänen et al., 2016), global magnetospheric simulations ( Tenfjord et al., 2015; Eggington et al., 2022) and auroral observations ( Fear and Milan, 2012; Kullen et al., 2015). However, the inferred estimates span from a few tens of minutes to several hours. In the first statistical studies reporting the interhemispheric asymmetry associated with the fast earthward flows, the IMF conditions were inferred by averaging the IMF over a 130-min time interval prior to a flow event ( Pitkänen et al., 2013; Pitkänen et al., 2017). The direction of the IMF B _y component typically varies little compared to that of IMF B _z , and also much shorter IMF averaging windows (e.g., 15 min) have found to be suitable ( Pitkänen et al., 2019; Chong et al., 2022). A prolonged constant nonzero IMF B _y direction may be a prerequisite for clear signatures of the asymmetry. Clear tail responses (e.g., Pitkänen et al., 2018) are not often observed in the data and the superposed epoch analysis of the dusk-dawn velocity in IMF B _y reversals by Case et al. (2020) support this. Furthermore, from ionospheric observations, we know that at least for the northward IMF with a dominant B _y component, the asymmetry in the two-cell ionospheric convection pattern increases with time ( Grocott and Milan, 2014), which indicates also an increase of the degree of twisting in the magnetotail.

Another open question is what is the role of the geomagnetic dipole tilt angle in the tail flow asymmetry. Indications that the ionosphere is influenced not only by IMF B _y but a combination of IMF B _y and the Earth’s dipole tilt has been reported long ago (e.g., Friis-Christensen and Wilhjelm, 1975; Crooker, 1992; Ruohoniemi and Greenwald, 2005). Recent statistical studies by Reistad et al. (2020), Holappa et al. (2021), Ohma et al. (2021) and Laitinen et al. (2024) indicate that the influence by both IMF B _y and the dipole tilt angle appears, e.g., in the ionospheric field-aligned current pattern, geomagnetic activity, substorm occurrence frequency, Hall conductance and the strength and width of the dawnside auroral electron precipitation region. This implies that the different ionospheric response during positive and negative IMF B _y in combination with different dipole tilt angles may be coupled to tail dynamics as well. In which way a combination of Earth dipole tilt and IMF B _y may affect the tail dynamics and topology, still needs to be examined. It is known though that the dipole tilt angle has a strong effect to the shape of the tail neutral sheet. For positive dipole tilts, the neutral sheet is displaced poleward of the so-called tail equatorial plane, with the neutral sheet flanks being curved below the equatorial plane, such that it has a warped shape (e.g., Tsyganenko et al., 2015). For negative dipole tilts the warping has an opposite sense. As discussed in Section 1, a nonzero IMF B _y in turn causes a rotation of the neutral sheet around the Sun-Earth axis (e.g., Tsyganenko et al., 2015). How different combinations of the dipole tilt and IMF B _y affect the tail plasma sheet convection is yet to be discovered.

Further research is needed to deepen our understanding of the solar wind-magnetosphere coupling.