National Institute of Polar Research, Tokyo, Japan
The ionospheric cusp/cleft are the focal points of geomagnetic field lines threading a wide area of magnetospheric boundary layers. Therefore auroral and electromagnetic phenomena observed through this small window are manifestations of physical processes working in the magnetospheric boundary layers. Field-aligned currents (FACs) are important physical quantities for the diagnosis of the magnetosphere, because they are generated when transverse momentum is transferred from one plasma regime to another along the magnetic field lines connecting the two. It is well known that the intensity and the spatial distribution pattern of the dayside FACs are controlled by the magnitude and the direction of the interplanetary magnetic field (IMF) (e.g., the review by Potemra , and references therein). This implies that the reconnection on the dayside magnetopause accounts for the major part of the solar wind-magnetosphere interaction. Namely, the solar wind momentum enters the magnetosphere through the region of open magnetic field lines after the magnetopause reconnection. Thus from the low-altitude observation of the cusp/cleft currents, we can infer the nature of the solar wind-magnetosphere interaction.
The purpose of this paper is to present the author's understanding of the dayside FAC systems in terms of momentum transfer from the magnetosphere to the ionosphere. In section 2, we first consider the basic role of FACs in magnetospheric dynamics and show the viewpoint in interpreting the observed FACs. The approach employed here is in the spirit of Southwood . In the subsequent two sections, we apply this idea to dayside large-scale FAC systems under particular IMF conditions. The first topic (section 3) is the FACs for periods of strong IMF By and southward Bz [Watanabe et al., 1996; hereafter referred to as paper 1], and the second topic (section 4) is the FAC systems in the magnetospheric ground state associated with extremely small IMF [Watanabe et al., 1998; hereafter referred to as paper 2]. Based on the results in papers 1 and 2, we discuss the three-dimensional current systems that establish the force balance between the hot magnetospheric plasma and the cold ionospheric plasma. Since details of the observations are already given in papers 1 and 2, as for the observational aspect, only a summary will be presented in this paper. And also, with the desire to keep the paper short, no attempt will be made to list complete references to previous work. We apologize to all the researchers in the field whose contributions are not acknowledged here. For the historical summary of the past studies, see Potemra , and references in papers 1 and 2.
The fundamental role of FACs is to transport transverse momentum, together with a transverse electric field and electromagnetic energy, along magnetic field lines from the magnetosphere to the ionosphere. In this section we picture the role of FACs in the spirit of Southwood . Figure 1 shows a simple magnetosphere-ionosphere coupling model in which two slabs of plasma are linked by magnetic field lines. The top slab represents the magnetosphere and the bottom slab the ionosphere. The whole region is immersed in a strong magnetic field B0. Suppose that the plasma in the top slab is moving at some initial time but the bottom slab is initially stationary. Since the two slabs are coupled by the magnetic field, the bottom slab will also start to move in time. Namely, there must be a general momentum transfer from the top slab to the bottom slab. This is easily pictured in the MHD (magnetohydrodynamic) description, as shown in Figure 1. In the top slab, the field lines are bent in order that there is a component of field line tension in the opposite direction to the plasma flow, while in the bottom slab, the field lines are bent so that the field line tension is the same direction as the flow in the top slab. In other words, the plasma stress in the top slab is transferred to the bottom slab in the form of Maxwell stress (electromagnetic momentum tensor: B0D B/m0 ). At the same time, a transverse electric field ( E ) and electromagnetic energy (Poynting flux: E D B/m0) are also transferred from the magnetosphere to the ionosphere.
From the observational point of view, it is useful to use a description involving electric currents rather than just Maxwell stress. In the current description, the momentum transfer along magnetic field lines is accomplished by a closed current circuit consisting of perpendicular currents ( J ) in the magnetosphere and the ionosphere, and the two oppositely-directed FACs ( J|) that connect the transverse currents. This current circuit is shown by dashed lines in Figure 1. In the top slab, the J B (i.e., ( B0 D B/m0) ) force is balanced by the force driving the magnetospheric plasma ( F: pressure gradient force, inertia force, or viscous force), and J E < 0 means that the plasma energy (thermal energy or kinetic energy) is converted to electromagnetic energy. In the bottom slab, the J B force drives the ionospheric plasma in the same direction as the magnetospheric plasma flow, and J E > 0 means that the electromagnetic energy is dissipated in the ionosphere.
Thus the role of FACs is summarized as follows: they transfer (1) a transverse electric field ( E), (2) transverse momentum ( B0DB/m0 per unit area per unit time), and (3) electromagnetic energy ( E DB/m0 per unit area per unit time). The main purpose of the FAC study is to infer the force F in the magnetosphere from the observation of J|. (It is very difficult to observe in situ F.) To this end, we must know which two FACs are coupled in the sense that they constitute the current circuit in Figure 1. Practically, the following two points are important for the interpretation of the data: (1) the balance of currents flowing into and away from the ionosphere (since the inflow and the outflow should be equal if the two FACs are coupled), and (2) the direction of E (since the ionospheric coupling current J must flow in the direction of E). Bearing these points in mind, in the next two sections we discuss the three-dimensional current systems under the two IMF conditions.
The discussion in this section is based on the results in paper 1. We investigated the By -dependent FAC systems on the dayside using DMSP F6 and F7 data. This topic has been investigated by many authors using polar-orbiting satellites (e.g., [Bythrow et al., 1988] (DMSP F7); [Erlandson et al., 1988] (Viking); [Taguchi et al., 1993] (DE 2); see also the review by Potemra ). By these authors, a consensus has been established regarding the By -dependence of the current polarity and the magnetospheric source regions of FACs: A longitudinally elongated pair of FACs occurs in the cusp/mantle region in the midday sector, and its flow direction changes systematically depending on the IMF By polarity. (The details of the By -dependent current system will be given below with our refined model.) In the past studies, however, the detailed spatial structure of the midday FAC systems has not been clarified, because all the models in the past are statistical patterns assembled from a large number of satellite observations during different days of various magnetospheric conditions. In paper 1, we performed not a statistical but an event-based study to determine a synoptic map of dayside FAC systems. We investigated several events of prolonged geomagnetic disturbance in which quasi-stationary IMF with |By| 4 nT and Bz < 0 persisted for more than one day. This approach enabled us to obtain synoptic maps of dayside plasma regimes and FACs.
Figure 2 shows a schematic model showing the relationship between plasma domains and FACs (adapted from Figure 7 of paper 1). This picture was drawn from several synoptic maps of plasma domains and FACs. (Four cases are presented by combining the northern and southern hemisphere maps; for example, "By < 0 north" means the northern hemisphere map for negative By periods.) We identified five plasma domains: "inner plasma sheet," "outer plasma sheet," "cusp," "mantle," and "cleft." The identification scheme is essentially the same as that by Newell and his coworkers [Newell and Meng, 1988; Newell et al., 1991a, 1991b], and these five plasma domains correspond approximately to their "central plasma sheet," "boundary plasma sheet," "cusp," "mantle," and "low-latitude boundary layer," respectively. Here we should mention that our cleft is not necessarily the same as the low-latitude boundary layer by Newell et al. [1991b]. We used the criterion for the "boundary layer" by Newell and Meng  in identifying the cleft. Therefore the cleft includes a region of plasma-sheet-like somewhat hard precipitation, which we infer is the ionospheric projection of the inner part of the flankside low-latitude boundary layer (the stagnation region [e.g., Traver et al., 1991]). In contrast, the low-latitude boundary layer by Newell et al. [1991b] is restricted to a more magnetosheath-like region.
The FACs that occur in the cusp/mantle region are controlled by IMF By. In the midday sector, the By -dependent current system consists of two current sheets elongated longitudinally. The equatorward current is roughly associated with the cusp precipitation and its location corresponds to the midday part of the region 1 "ring." Here we call it the "midday region 1." The poleward current is roughly associated with the mantle precipitation. Historically, a variety of names have been given to this poleward current. For example, Iijima and Potemra  originally called it the "cusp region" current (simultaneous particle data were not available in those days). In order to avoid confusion, here we call it "region 0." We generally use the term region 0 for the FACs poleward of region 1, even if they are located away from noon. For By < 0 north and for By > 0 south (Figure 2a), the region 0 current flows into the ionosphere, while the midday region 1 current flows away from the ionosphere. For By > 0 north and for By < 0 south (Figure 2b), the flow directions of region 0 and the midday region 1 are totally reversed. In contrast, the flow directions of FACs found in the cleft and the plasma sheet are independent of By. The pair of FACs in the noon sector drives eastward or westward ionospheric convection depending on By. For By < 0 north and for By > 0 south (Figure 2a), the plasma flow at noon is eastward, while it is westward for By > 0 north and for By < 0 south (Figure 2b).
As schematically shown in Figure 2, the spatial distribution pattern of plasma domains for By > 0 north and for By < 0 south (Figure 2b) is essentially the mirror image of that for By < 0 north and for By > 0 south (Figure 2a) with respect to the noon meridian, although in fact minor dawn-dusk asymmetries do exist. In order to treat the four cases synthetically, here we introduce a new terminology. We call the dawnside in Figure 2a and the duskside in Figure 2b the "upstreamside," because they are located upstream of the midday plasma convection. Similarly, we call the duskside in Figure 2a and the dawnside in Figure 2b the "downstreamside." If we use the terms upstreamside and downstreamside, the By -dependent characteristics of plasma regimes are synthetically summarized as follows: (1) The cusp shifts toward the upstreamside. (2) The mantle extends on the downstreamside, while it is absent on the upstreamside. (3) The upstreamside cleft is confined in a narrow region, while the downstreamside cleft occurs in a wide range of magnetic local time (MLT). (4) The upstreamside outer plasma sheet extends poleward or sunward. Of the four plasma regime characteristics, the second and the third are important in the discussion on the FACs below.
As noted in section 2, the imbalance between FACs flowing into and away from the ionosphere provides key information to identify the coupling of two FACs. For southward Bz periods, FACs flow in approximately east-west-aligned sheets, which assures us of the validity of the infinite current sheet approximation in interpreting FACs from a single polar-orbiting satellite. Namely, the FAC density is given by the gradient of the cross-track component of the transverse magnetic field disturbance ( DB ), and the intensity of the sheet current is given by DB/m0. With this assumption, we investigated the current imbalance (1) between region 0 and the midday region 1, and (2) between the upstreamside cleft region 1 and the downstreamside cleft region 1. Figure 3 shows a schematic representation of the results in paper 1 (the satellite view for the By < 0 north case). The number of arrows qualitatively represents the magnitude of the total current.
We first investigated the latitudinal imbalance of the region 0 and midday region 1 currents in the noon sector. We examined the FAC intensity (the latitudinally integrated current density given in A m -1 ) of region 0 and the midday region 1 when DMSP F7 crossed the cusp/mantle region in the north-south direction. Semi-statistical results (Figure 9 of paper 1) showed that (1) for about half of the cusp/mantle crossings, the region 0 intensity was well balanced with the midday region 1 intensity, and (2) for the other half of the cusp/mantle crossings, the region 0 intensity exceeded the midday region 1 intensity. (We never found the cases in which the midday region 1 intensity significantly exceeded the region 0 intensity.) The former fact indicates that the region 0 and midday region 1 currents are coupled in the ionosphere and therefore also in the magnetosphere (see Figure 1). The latter fact, however, implies that the coupling is not perfect and another mode of FACs does exist. It should be remembered here that the region 0 associated with the mantle precipitation is observed in a wide local time range extending toward the downstreamside as shown in Figure 2, while the midday region 1 is found only near noon. Thus we conclude that the total current of region 0 is greater than that of the midday region 1 (see Figure 3). Namely, there exists an extra part of region 0 that is not connected to the midday region 1 in the magnetosphere.
We now need to consider which current the extra part of region 0 is connected to in the magnetosphere. We then examined the total region 1 current (the longitudinally integrated current intensity given in A) associated with the clefts. We found that the total current of the downstreamside cleft region 1 always exceeded that of the upstreamside cleft region 1 (Table 3 of paper 1). It should be remembered that the upstreamside cleft is confined in a narrow region, while the downstreamside cleft is found in a wide range of MLT. Not only the current intensity but also its longitudinal scale contributes to the excess of the downstreamside cleft region 1. This imbalance in the cleft currents are schematically shown in Figure 3. It should be noted here that the flow direction of region 0 is opposite to that of the downstreamside cleft region 1, while it is the same as that of the upstreamside cleft region 1. This indicates that the extra part of region 0 is connected to the downstreamside cleft region 1 in the magnetosphere. This coupling is consistent with the direction of the transverse electric field ( E ). That is, the coupling current ( J ) in the magnetosphere flows in the dusk-to-dawn direction (whereas E in dawn-to-dusk).
Thus the balance/imbalance of dayside FACs associated with the cusp/mantle/cleft regions indicate that there are at least two modes of FAC coupling: One is the coupling of region 0 and the midday region 1, and the other is the coupling of region 0 and the downstreamside cleft region 1. Figure 4 (adapted from Figure 10 of paper 1) depicts the two coupling modes working for periods of strong By and southward Bz. (The By < 0 north case is given here.)
Figure 4a shows the coupling of region 0 and the midday region 1 in the cusp/mantle region near the Earth ( < 10 RE). For periods of By < 0, the cusp shifts dawnward in the northern hemisphere. The magnetosheath plasma enters the cusp from the dawnside edge and flows duskward. If the force ( F ) acting on the plasma is directed duskward (i.e., deceleration of the plasma flow or a dawnward pressure gradient), which is a reasonable assumption in this situation, the duskward momentum is transmitted to the ionosphere in the form of Maxwell stress. In the current description, J in the source region flows so that F+ J B=0 , and it generates a pair of FACs (i.e., region 0 and the midday region 1). It is worth remarking here that the process depicted in Figure 4a does not occur just after the reconnection on the dayside magnetopause. The magnetic tension in the source region should be dawnward (cf. Figure 1), whereas it is duskward just after the dayside reconnection.
Figure 4b shows the coupling of region 0 and the downstreamside (duskside for the By< 0 north case) cleft region 1 in the flank of the magnetosphere. In this coupling process, the tailward inertia force would be the source of momentum. If the low-latitude boundary layer is totally closed, the coupling of region 0 and region 1 must occur at the juncture of the high- and low-latitude boundary layers. Such a coupling is questionable, since it is unclear how the Maxwell stress depicted in Figure 1 is maintained on the field lines threading the juncture. However, during southward Bz periods, the low-latitude boundary layer, in particular its outer part, is suggested to be open [e.g., Mitchell et al., 1987]. Therefore the coupling in Figure 4b is quite reasonable if we consider a dynamo on open field lines.
The topic of this section is the FAC systems in the extremely quiet magnetosphere associated with very small IMF, or the ground state of the magnetosphere. The discussion is based on the results in paper 2. The IMF control of FACs [e.g., Potemra, 1994] and plasma convection patterns [e.g., Heppner and Maynard, 1987] in the polar ionosphere suggests that the magnetic reconnection on the dayside magnetopause accounts for the major part of the solar wind-magnetosphere interaction. According to the anti-parallel reconnection model, the momentum transfer from the solar wind to the magnetosphere is expected to be minimized when By=Bz=0. Thus, as a continuation of paper 1 (section 3), this work was motivated by the question what FACs remain in the quiet magnetosphere associated with By=Bz=0. To date, only a few studies have been made on the quiet-time FACs [Rich and Gussenhoven, 1987; Hoffman et al., 1988], although the investigation of the baseline magnetosphere has been the focus of much research effort for years (the article by Gussenhoven  gives a good review of the observational aspect). The objective of this study is to determine the FAC systems in the baseline magnetosphere associated with very small IMF and to reveal a basic mode of the solar wind-magnetosphere interaction that is concealed by IMF changes and substorm activities in the usual magnetospheric state. We defined the ground state of the magnetosphere by the following criterion: (1) Kp = 0, (2) |By| 1.5 nT, and (3) -0.5 nT Bz 1.5 nT. Using Magsat or DMSP (F7/F6) satellite data, we investigated four days (events) of prolonged ( > 12 h) geomagnetic quiescence that satisfied the above criterion. For all the four events, the total magnitude of the IMF was also very small (mostly in the range of 2.5-3.5 nT). As in paper 1, we performed an event-based study for the four events to determine synoptic maps of plasma regimes.
Figure 5 shows the observational summary adapted from Figure 9 of paper 2. The figure was drawn from a pair of northern and southern synoptic maps obtained in one of the four events, but the various features discussed below were common to all the four events. From the precipitating particles, we have identified three plasma domains: the inner plasma sheet, the cusp, and the "boundary plasma region (BPR)." At first we tried to identify five plasma domains as in paper 1 (section 3). However, the precipitation from the mantle, the cleft, and the outer plasma sheet showed similar characteristics of energy spectra during the events examined, and we found it difficult to discriminate these three plasma domains. Therefore we treated them as one plasma regime and called it the BPR. As shown in Figure 5, the BPR occurs poleward of the inner plasma sheet and encircles the polar cap. A noteworthy feature of the BPR in the ground state is that its latitudinal width is very thin in the premidnight sector but rather thick on the flanksides, which would be related to the magnetotail structure in the ground state. In contrast, the cusp does exist even during the ground state and is unambiguously identified near noon embedded in the BPR.
The major FAC systems that persist in the ground state are the prenoon and the postnoon region 1, and the prenoon and the postnoon region 0 ( Iijima and Potemra's  cusp region currents). Although the basic pattern by Iijima and Potemra (i.e., region 1, region 2, and the dayside region 0) is still discernible even in the ground state, the well-defined region 1 and region 2 are no longer observed on the flanksides and on the nightside. They are disintegrated into small-scale weak FACs. In addition, there are no significant large-scale FACs in the cusp. Thus the elements of the ground state FACs are the prenoon and the postnoon region 0/region 1 pair. These features are emphasized in Figure 5. Figure 5 does not mean that there are no FACs at other local times than the prenoon and postnoon sectors. Both region 0 and region 1 currents occur on the field lines threading the BPR.
The prenoon and the postnoon region 0/region 1 occur in pairs. From the DMSP F7 observations, we found the following characteristics of the prenoon region 0/region 1 pair at 0900-1000 MLT (Plate 1 and Table 2 of paper 2): (1) The two currents are well approximated by a pair of current sheets elongated longitudinally (see Figure 5). (2) The intensities of region 0 and region 1 are perfectly balanced with each other with a typical value of 140 nT (0.11 A m -1 ). These facts indicate that region 0 and region 1 in the prenoon sector are coupled in the ionosphere via Pedersen currents, meaning that they are also coupled in the magnetosphere (see Figure 1). Although we do not have sufficient data in the postnoon sector, we infer that similar characteristics are also found for the postnoon region 0/region 1 pair. The couplings of region 0 and region 1 in the prenoon and postnoon sectors are suggested to be the fundamental mode of FACs that persists in the magnetospheric ground state.
From the flow directions of region 0 currents, one may expect the presence of another coupling mode - the coupling of the prenoon and the postnoon region 0. However, this coupling is denied from the energetics. The overall D B observations in the polar region indicate that the ionospheric convection in the ground state is basically the two-cell pattern, and that on the noon meridian it is weakly antisunward (Figure 3 of paper 1). That is, the electric field in the polar cap is still dawn-to-dusk in the ground state. The antisunward convection in the polar cap occurs because the region 0 intensity never exceeds the concurrent region 1 intensity. This feature of region 0 is contrasted with that of the NBZ currents [Iijima et al., 1984] for strongly northward IMF (region 0 in a broad sense), which are found to extend over the noon meridian with the intensity greater than the region 1 intensity. In the case of the ground state, if we connect the prenoon and the postnoon region 0, the perpendicular coupling currents ( J) flow dusk-to-dawn in the ionosphere and dawn-to-dusk in the magnetosphere. This is inconsistent with the energetics, because it follows that J E < 0 (a dynamo) in the ionosphere and J E > 0 (a dissipater) in the magnetosphere. Thus we cannot connect the prenoon and the postnoon region 0. Instead, the coupling should occur between the prenoon region 0 and the prenoon region 1, and between the postnoon region 0 and the postnoon region 1.
The region 0/region 1 current system is found on field lines whose magnetospheric mapping is uncertain. Hence a problem remains as to where in the magnetosphere the region 0/region 1 coupling occurs. There are three ways of thinking: (1) the region 0/region 1 pair is mapped into a closed boundary region such as the low-latitude boundary layer (Figure 6a); (2) it is mapped into an open boundary region such as the high-latitude boundary layer (Figure 6b); and (3) it is mapped into the juncture of a closed and an open boundary region (Figure 6c). At present we cannot say with certainty which is correct, but we can comment some implications.
The third model (Figure 6c) shows dynamos at the juncture of a closed and an open boundary region. The problem of this model is its magnetic field configuration. In actuality, there exists no flux tube (with a finite cross section) threading the juncture itself; therefore this model should be regarded as a dynamo stretching over both the closed and open regions. On the closed region side, the coupling current ( J) flows near the equatorial plane. In contrast, the coupling current on the open region side would flow in the region away from the equator, since the field lines are extending down to the magnetotail. The Maxwell stress (i.e., the J B force) should be continuous from the closed region to the open region. It is questionable that such a configuration can be materialized by the field lines threading the region in the vicinity of the open-closed juncture. We think this model is unlikely.
The second model (Figure 6b) shows dynamos in an open boundary region. Since open field lines are connected to the solar wind, they are removed to the magnetotail in the end. In order for this dynamo process to work continuously, open magnetic field lines must be supplied by the closed-to-open flux transfer on the dayside magnetopause. Similarly the open-to-closed flux transfer in the magnetotail must occur for the entire magnetospheric system to be stationary. This could be possible if reconnection occurs constantly even during the ground state. However, from the precipitating particle characteristics of the BPR, we think this model is also unlikely. The equatorward half of the BPR (the domain of region 1) always showed cleft-like hard precipitation suggesting a closed boundary region (Plate 1 of paper 2). It is, therefore, unthinkable that the entire region 1 is on open magnetic field lines. The balance of region 0 and region 1 in the ground state implies that almost all the region 1 currents in the prenoon and postnoon sectors must be coupled with the region 0 currents. Accordingly the coupling of region 0 and region 1 on open field lines is not the case for the ground state. It should be remarked here, however, that this coupling is possible for Bz< 0 periods. For southward IMF, the region 1 intensity is much greater than the region 0 intensity. In addition, the outer part of the low-latitude boundary layer is suggested to be open for southward IMF [e.g., Mitchell et al., 1987]. Hence, in contrast to the ground state, it is possible that a part of region 1 is coupled with region 0 on open field lines.
The first model (Figure 6a) shows dynamos in a closed boundary region. This model is consistent with the well-known viscous interaction between the solar wind and the magnetosphere in the low-latitude boundary layer, and we think the most reasonable model of the three. For southward IMF, it has been suggested that the non-cusp region 1 in the prenoon and postnoon sectors is mapped into the low-latitude boundary layer, while region 0 is projected into the high-latitude boundary layer (mantle). Hence a problem still remains as to why, under this special magnetospheric condition, region 0 is included in the closed region and how the transition occurs. For this problem, the model by Siscoe et al.  is very suggestive. They developed an analytic model that joins the high- and low-latitude boundary layers with the ionosphere by an electric current circuit. Their results show that when the polar cap potential is less than 35 kV, region 0 currents appear within the region of closed magnetic field lines. The transition occurs when the voltage in the high-latitude boundary layer becomes less than the critical value controlled by the voltage in the low-latitude boundary layer.
Finally we would like to mention one important characteristic of region 1 currents. The intensity of the dayside region 1 currents in the ground state shows seasonal variations. The intensities in the winter hemisphere were significantly smaller than those in other seasons (see paper 2), which is the same characteristic found by Fujii and Iijima . This manifests that the region 1 intensity on the dayside is controlled by ionospheric conductivities, indicating that the region 1 current is driven by a voltage generator in the magnetosphere rather than by a current generator. The cross-polar potential in the ground state estimated from the region 1 intensity (140 nT) was on the order of 10-20 kV (see paper 2). This suggests that a voltage generator of 10 kV is always operating in the dawn and dusk low-latitude boundary layers. The coupling of region 0 and region 1 in the ground state substantiates the existence of the voltage generator in the closed boundary region.
The dynamical nature of FACs is that they transfer transverse momentum along magnetic field lines from one plasma to another, which is easily pictured in the MHD description. Thus one important aspect of the FAC study is to infer the momentum acting on the source region plasma from the observed FACs. In the electric current description, the momentum transfer is interpreted by a closed current circuit connecting the magnetosphere and the ionosphere (Figure 1). Therefore the first step of the FAC study is to identify which two FACs are coupled. We have applied this idea to dayside large-scale FAC systems under two particular IMF conditions: (1) |By| 4 nT, and Bz < 0, and (2) |By| 1.5 nT, and - 0.5 nT Bz 1.5 nT (and Kp=0 ). The latter concerns the extremely quiet magnetosphere associated with very small IMF (the ground state of the magnetosphere). For periods of strong By (IMF condition 1), we suggest at least two modes of FACs are working. One is the coupling of region 0 and the midday region 1 in the high-latitude near-Earth ( < 10 RE ) cusp (Figure 4a). The other is the coupling of region 0 and the downstreamside cleft region 1 in the flankside magnetospheric boundary layers (Figure 4b). Both modes are suggested to be operating on open magnetic field lines. For periods of very small IMF (IMF condition 2), one fundamental mode of FACs persists: the coupling of region 0 and region 1 on the flanksides of the magnetosphere. This coupling is inferred to originate in the closed low-latitude boundary layer (Figure 6a), presumably in consequence of the viscous interaction between the solar wind and the magnetosphere.
Bythrow, P. F., T. A. Potemra, R. E. Erlandson, L. J. Zanetti, and D. M. Klumper, Birkeland currents and charged particles in the high-latitude prenoon region: A new interpretation, J. Geophys. Res., 93 (A9), 9791, 1988.
Erlandson, R. E., L. J. Zanetti, T. A. Potemra, P. F. Bythrow, and R. Lundin, IMF By dependence of region 1 Birkeland currents near noon, J. Geophys. Res., 93 (A9), 9804, 1988.
Fujii, R., and T. Iijima, Control of the ionospheric conductivities on large-scale Birkeland current intensities under geomagnetic quiet conditions, J. Geophys. Res., 92 (A5), 4505, 1987.
Gussenhoven, M. S., Low-altitude convection, precipitation, and current patterns in the baseline magnetosphere, Rev. Geophys., 26 (4), 792, 1988.
Heppner, J. P., and N. C. Maynard, Empirical high-latitude electric field models, J. Geophys. Res., 92 (A5), 4467, 1987.
Hoffman, R. A., M. Sugiura, N. C. Maynard, R. M. Candey, J. D. Craven, and L. A. Frank, Electromagnetic patterns in the polar region during periods of extreme magnetic quiescence, J. Geophys. Res., 93 (A12), 14,515, 1988.
Iijima, T., and T. A. Potemra, Field-aligned currents in the dayside cusp observed by Triad, J. Geophys. Res., 81 (34), 5971, 1976.
Iijima, T., T. A. Potemra, L. J. Zanetti, and P. F. Bythrow, Large-scale Birkeland currents in the dayside polar region during strongly northward IMF: A new Birkeland current system, J. Geophys. Res., 89 (A9), 7441, 1984.
Mitchell, D. G., F. Kutchko, D. J. Williams, T. E. Eastman, L. A. Frank, and C. T. Russell, An extended study of the low-latitude boundary layer on the dawn and dusk flanks of the magnetosphere, J. Geophys. Res., 92 (A7), 7394, 1987.
Newell, P. T., and C.-I. Meng, The cusp and the cleft/boundary layer: Low-altitude identification and statistical local time variation, J. Geophys. Res., 93 (A12), 14,549, 1988.
Newell, P. T., W. J. Burke, C.-I. Meng, E. R. Sánchez, and M. E. Greenspan, Identification and observation of the plasma mantle at low altitude, J. Geophys. Res., 96 (A1), 35, 1991a.
Newell, P. T., W. J. Burke, E. R. Sánchez, C.-I. Meng, M. E. Greenspan, and C. R. Clauer, The low-latitude boundary layer and the boundary plasma sheet at low altitude: Prenoon precipitation regions and convection reversal boundaries, J. Geophys. Res., 96 (A12), 21,013, 1991b.
Potemra, T. A., Sources of large-scale Birkeland currents, in Physical Signatures of Magnetospheric Boundary Layer Processes, edited by J. A. Holtet and A. Egeland, pp. 3-27, Kluwer Academic Publishers, 1994.
Rich, F. J., and M. S. Gussenhoven, The absence of region 1/region 2 field-aligned currents during prolonged quiet times, Geophys. Res. Lett., 14 (7), 689, 1987.
Siscoe, G. L., W. Lotko, and B. U. Ö. Sonnerup, A high-latitude, low-latitude boundary layer model of the convection current system, J. Geophys. Res., 96 (A3), 3487, 1991.
Southwood, D. J., An introduction to magnetospheric MHD, in Solar System Magnetic Fields, edited by E. R. Priest, pp. 25-36, D. Reidel Publishing Company, 1985.
Taguchi, S., M. Sugiura, J. D. Winningham, and J. A. Slavin, Characterization of the IMF By -dependent field-aligned currents in the cleft region based on DE 2 observations, J. Geophys. Res., 98 (A2), 1393, 1993.
Traver, D. P., D. G. Mitchell, D. J. Williams, L. A. Frank, and C. Y. Huang, Two encounters with the flank low-latitude boundary layer: Further evidence for closed field topology and investigation of internal structure, J. Geophys. Res., 96 (A12), 21,025, 1991.
Watanabe, M., T. Iijima, and F. J. Rich, Synthesis models of dayside field-aligned currents for strong interplanetary magnetic field By, J. Geophys. Res., 101 (A6), 13,303, 1996.
Watanabe, M., T. Iijima, M. Nakagawa, T. A. Potemra, L. J. Zanetti, S.-I. Ohtani, and P. T. Newell, Field-aligned current systems in the magnetospheric ground state, J. Geophys. Res., 103 (A4), 6853, 1998.