Field‐Aligned Currents in Saturn's Nightside Magnetosphere: Subcorotation and Planetary Period Oscillation Components During Northern Spring

We newly analyze Cassini magnetic field data from the 2012/2013 Saturn northern spring interval of highly inclined orbits and compare them with similar data from late southern summer in 2008, thus providing unique information on the seasonality of the currents that couple momentum between Saturn's ionosphere and magnetosphere. Inferred meridional ionospheric currents in both cases consist of a steady component related to plasma subcorotation, together with the rotating current systems of the northern and southern planetary period oscillations (PPOs). Subcorotation currents during the two intervals show opposite north‐south polar region asymmetries, with strong equatorward currents flowing in the summer hemispheres but only weak currents flowing to within a few degrees of the open‐closed boundary (OCB) in the winter hemispheres, inferred due to weak polar ionospheric conductivities. Currents peak at ~1 MA rad−1 in both hemispheres just equatorward of the open‐closed boundary, associated with total downward polar currents ~6 MA, then fall across the narrow auroral upward current region to small values at subauroral latitudes. PPO‐related currents have a similar form in both summer and winter with principal upward and downward field‐aligned currents peaking at ~1.25 MA rad−1 being essentially collocated with the auroral upward current and approximately equal in strength. Though northern and southern PPO currents were approximately equal during both intervals, the currents in both hemispheres were dual modulated by both systems during 2012/2013, with approximately half the main current closing in the opposite ionosphere and half cross field in the magnetosphere, while only the northern hemisphere currents were similarly dual modulated in 2008.


Introduction
The large-scale electric currents that flow between the ionosphere and magnetosphere in planetary plasma systems are a fundamental feature associated with the transfer of momentum between these regions (e.g., Cowley, 2000;Hunt et al., 2018). At Saturn, as the subject of the present paper, two current systems have been identified via their associated perturbation magnetic fields and have been extensively studied using data from the Cassini spacecraft Hunt et al., 2014Hunt et al., , 2015Hunt et al., , 2016Southwood & Kivelson, 2009;Talboys et al., 2011;Talboys, Arridge, Bunce, Coates, Cowley, Dougherty, & Khurana, 2009). The first of these is the quasi-steady currents associated with subcorotation of the magnetospheric plasma relative to the neutral upper atmospheric flow, resulting from charged particle pickup and outward radial transport from the water gas source at the moon Enceladus on closed field lines, together with effects associated with the solar wind interaction on open field lines (e.g., Hill, 1979;Isbell et al., 1984;Cowley & Bunce, 2003;Cowley, Bunce, & Prangé, 2004;Pontius & Hill, 2009;Bagenal & Delamere, 2011). The second is the current systems associated with the planetary period oscillation (PPO) phenomenon, related to twin-vortex flows in the northern and southern polar ionospheres, which rotate around the planetary axis with close but separate periods near to~10.7 hr, whose oscillatory effects are ubiquitous throughout Saturn's magnetosphere Carbary & Mitchell, 2013;Espinosa & Dougherty, 2000;Gurnett et al., 2009;Hunt et al., 2014;Southwood & Cowley, 2014). In principle, a third current system may also be present associated with solar wind driving of Earth-like magnetospheric flow, which is fixed in local time (LT) and oppositely directed at dawn and dusk (Cowley, Bunce, & O'Rourke, 2004;Jackman & Cowley, 2006;Southwood & Kivelson, 2009). However, no such effect has been demonstrated in Cassini field data to date

Subcorotation and PPO Current Systems
The principal features of the subcorotation and PPO-related current systems at Saturn discerned from the recent studies cited above are sketched in Figure 1, adapted from Hunt et al. (2015), where Figures 1a-1c show the subcorotation current system, while Figures 1d-1f and 1g-1i show the northern and southern PPO-related currents, respectively. In this figure the currents are indicated by green arrowed lines and symbols and the magnetic perturbation fields by blue arrowed lines and symbols, where circled dots and crosses indicate vectors pointing out of or into the plane of the figure, respectively. Figures 1a, 1d, and 1g show the ionospheric currents in a view looking down from the north, "through" the planet in the case of the southern PPO system in Figure 1g. Arrowed red dashed lines show the atmospheric/ionospheric flows associated with each system, assuming that the PPOs are driven outward from the planetary atmosphere in Figures 1d and  1g. Figures 1b, 1e, and 1h show the currents and perturbation fields in a magnetic meridian plane (the principal meridian for the PPO currents as discussed further below), where the black arrowed lines show the background magnetic field, approximately symmetric about Saturn's rotation axis (Burton et al., 2010). For future reference, Figures 1c, 1f, and 1i show sketches in the same meridian planes of the azimuthal field perturbations just above the ionosphere, B φi , for both the northern (N) and southern (S) hemispheres. The central vertical arrowed lines mark the position of the poles, with colatitude from the poles increasing on either side, θ iN in the north and θ iS in the south. Vertical dotted lines indicate the location of the open-closed field line boundary (OCB) (Gurnett et al., 2010;Jinks et al., 2014).

Subcorotation Current System
Given the polarity of Saturn's internal field, opposite to Earth, the subcorotation current system in Figures 1a-1c consists of a distributed current flowing downward over the subcorotating polar regions, closed by a ring of upward current flowing at lower latitudes in each hemisphere where the plasma angular velocity increases toward that of the planet, Ω S Cowley & Bunce, 2003;Cowley, Bunce, & O'Rourke, 2004;Hill, 1979). Note that throughout this paper "upward" and "downward" are with respect to the planet and its upper atmosphere/ionosphere, that is, approximately radially outward and inward, respectively. The torque associated with the j × B force of the related equatorward directed ionospheric Pedersen current (Figure 1a) balances the frictional torque due to ion-neutral collisions in the ionosphere, with this torque being transferred to the equatorial plasma via the closure currents directed radially outward in the magnetosphere. The associated magnetic perturbations (Figure 1b) result in magnetospheric field lines being bent out of meridian planes into a "lagging" configuration, with negative azimuthal fields in the northern hemisphere and positive azimuthal fields in the southern hemisphere. Dynamics resulting in supercorotation of the magnetospheric plasma would correspondingly produce a "leading" configuration with reversed azimuthal fields, currents, and angular momentum transfer. Given subcorotation, the azimuthal field just above the ionosphere ( Figure 1c) increases monotonically from zero at each pole due to the increase in atmospheric velocity with distance from the rotation axis, then more rapidly across the OCB associated with an increase in ionospheric conductivity due to hot plasma precipitation on closed field lines, and then returns near to zero across the ring of upward directed current at lower latitudes (e.g., Hunt et al., 2014Hunt et al., , 2015. In the equatorial plane the upward field-aligned current region extends inward to radial distances of~10 R S (where R S is Saturn's 1-bar equatorial radius equal to 60,268 km). In Figures 1b and 1c these perturbations are shown for simplicity as being of equal magnitude in the northern and southern hemispheres, though as we will see in this paper, significant differences occur related to Saturn's seasons.

PPO Current Systems
By comparison, the principal currents of the two PPO systems flow into the ionosphere on one side of the planet and out on the other side (Figures 1d, 1e, 1g, and 1h), closing partly across the corresponding polar ionosphere, partly through the near-equatorial magnetosphere, and partly via the opposite ionosphere, where the cross-field currents are again associated with a j × B force on the plasma Southwood & Cowley, 2014). These systems generate a quasi-uniform perturbation field in the nearequatorial region interior to the field-aligned currents, directed outward in both Figures 1e and 1h, which weakens in the inner regions of the magnetosphere across the interior closure currents. This perturbation field closes principally over the corresponding pole, north for the northern system in Figure 1e and south for southern system in Figure 1h, extending also into the equatorial region, as shown by the fields directed into the plane in these figures. In these regions the perturbation takes the form of a quasi-dipolar field with the effective dipole being transverse to the spin axis and directed out of the diagrams in Figures 1e and 1h. However, these closure fields have been found not to extend with measureable amplitude over the "opposite" poles (Andrews et al., 2012;Hunt et al., 2015). These features are reflected in the ionospheric perturbation fields sketched in Figures 1f and 1i, where the apparent discontinuity in field value at the pole is due only to the switch in sense of the azimuthal unit vector across this axis. We note that in the equatorial region the transition of the perturbation field from quasi-uniform to quasi-dipolar, associated with a reversal in sense of the azimuthal field component, takes place at radial distances of~15-20 R S , while the main PPO currents extend inward to~10 R S (similar to the subcorotation upward current), and the closure currents to inside of~5 R S (Andrews, Coates, et al., 2010;Hunt et al., 2014Hunt et al., , 2015. For comparison, a typical stand-off distance of the subsolar magnetopause is~25 R S (e.g., Kanani et al., 2010;Pilkington et al., 2015).
Azimuthal position with respect to the PPO systems is defined through the phase functions Ψ N for the northern system and Ψ S for the southern system, as shown in the polar views in Figures 1d and 1g. These are defined for both systems such that the equatorial quasi-uniform perturbation field points radially outward at Ψ N,S = 0°. Figures 1e and 1h then show cuts in the Ψ N,S = 90°À 270°meridian plane, as marked in those figures. The radial fields thus vary throughout the domain as~cos Ψ N,S , while the azimuthal fields vary as sin Ψ N,S in the quasi-uniform field region, in lagging quadrature with the radial field, and as~À sin Ψ N,S in the quasi-dipolar field region, in leading quadrature with the radial field. Colatitudinal perturbation fields are also present joining the near-equatorial and polar perturbation field regions, which vary as~À cos Ψ N for the northern system, in antiphase with the radial field, and~cos Ψ S for the southern system, in phase with the radial field. These two systems then rotate about the planetary axis with (in general) separate periods close to that of the planetary rotation period, such that Ψ N,S increase monotonically with time at a fixed position. As discussed further in section 1.3, the periods of the two systems vary by~±1% about~10.7 hr over Saturn's seasons, compared with a value of~10.57 hr for Saturn's deep rotation period estimated by Read et al. (2009) from atmospheric dynamics considerations.

Cassini Measurements of Field-Aligned Current Signatures
As is clear from Figure 1, field-aligned current signatures can be determined from latitude profiles of the associated azimuthal magnetic perturbation fields on spacecraft passes over the planet's polar regions. Passes at smaller radial distances are preferred due to the larger amplitudes of the perturbation fields and the rapidity of crossing the current layers resulting in near "snapshots" being obtained of the perturbations at fixed time. Such passes also penetrate to field lines located deeper within the inner magnetosphere, thus more fully characterizing the current system. As noted in section 1.1, the principal currents of both subcorotation and PPO systems extend inward to~10 R S in the equatorial plane, such that field-aligned current studies at Saturn require sufficiently inclined orbits with near-periapsis equatorial crossings within~10 R S to provide adequate coverage of the main currents.
Properties of Cassini's orbit around Saturn are shown in Figure 2 over an 11-year interval from the start of 2005 to the end of 2015, in order to identify intervals during which such studies are feasible. Time in days is shown at the bottom of the figure, where t = 0 corresponds to 00 UT on 1 January 2004, while year boundaries are shown in red at the top of the figure. The times of Cassini periapsides are shown by black tick marks at the top of figure and marked with revolution (Rev) orbit numbers, defined from apoapsis to apoapsis, every 10 Revs. Saturn vernal equinox is marked by the vertical red dashed line, separating postsolstice southern summer conditions from presolstice northern spring (see section 1.3). Figure 2a shows the latitude range on each Rev (black lines), plotted at the time of periapsis, while the red and blue circles show the latitude of apoapsis and periapsis, respectively. Three principal intervals of highly inclined orbits are evident, the first spanning 2006/2007, the second 2008/2009, and the third 2012/2015. Data from the first of these intervals were studied by , Kivelson (2009), andHunt et al. (2016), while data from the second were studied by Talboys, Arridge, Bunce, Coates, Cowley, Dougherty, and Khurana (2009), Talboys et al. (2011), and Hunt et al. (2014, both intervals corresponding to late southern summer. Here we newly concentrate on data from the third interval, corresponding to northern spring. Figure 2b shows the radial range of each Rev on a log scale, where the radial distance of periapsis is of relevance for reasons given above. During the first interval of inclined orbits periapsis was located well below~10 R S at the beginning and end of the interval but rose above this limit in the central part. Consequently, useful field-aligned current data were obtained in two intervals spanning Revs 30-36 during late 2006 and Revs 41-44 in early 2007, as shown by the vertical blue bars (e.g., Hunt et al., 2016). In the second interval, periapsis was sufficiently low during the early part of the interval spanning Revs 59-95 in 2008 (Hunt et al., 2014(Hunt et al., , 2015, as again shown by the blue bar, but subsequently also became too large. A similar pattern occurred during the third interval of highly inclined orbits, with sufficiently small periapsides occurring for Revs 168-194 in 2012/2013 as examined here, again marked by the blue bar, before also becoming too large. We note, however, that the periapsis distances on the 2012/2013 Revs, typically~6-8 R S , are rather larger than those of the 2008 prime comparator data set, typically~3-5 R S , such that by comparison the 2012/2013 passes are slower and penetrate less deeply into the inner magnetosphere. Figure 2c shows the LT of apoapsis (red) and periapsis (blue) on each Rev. The LT of periapsis gives an initial indication of the LT of the spacecraft passes over the field-aligned current regions. During the late 2006 interval this was located near noon, moving sharply to predawn in early 2007, and then more slowly to the midnight sector in the 2008 interval. These data allowed Hunt et al. (2016) to compare southern hemisphere current signatures between midnight and dawn/noon LTs under late southern summer conditions. During the third interval periapsis was located in the postmidnight/predawn sector, similar to the second interval, thus allowing comparison of nightside currents under late southern summer and northern spring conditions as discussed here. Figure 2d shows the orbit period of each Rev from apoapsis to apoapsis, plotted at the time of periapsis. During the third interval studied here, the period was initially~20 days but subsequently decreased tõ 10 days near the end of 2012, comparable to that during the 2008 interval. Figure 3 shows physical parameters relevant to this study, plotted over the same interval as Figure 2. Figure 3a shows the latitude of the Sun at Saturn, indicative of Saturn's seasons, which passed through zero from southern summer to northern spring at vernal equinox in mid-2009. The prior southern solstice was in late 2002, and the following northern solstice in mid-2017. During the 2012/2013 interval studied here, again marked by the blue bar, the solar latitude increased from +15°to +19°, such that the northern atmosphere to~17°northern colatitude was in permanent sunlight, while the southern ionosphere to~17°southern colatitude was in permanent darkness. The opposite asymmetry occurred during the 2008 interval during which the solar latitude varied from À8°to À4°, such that the southern atmosphere to~6°colatitude was in permanent sunlight, while the northern atmosphere to~6°colatitude was in permanent darkness.

Physical Conditions
Figures 3b-3e show parameters of the northern (blue) and southern (red) PPO systems. Figure 3b shows the PPO periods, τ N,S , where values from 2005 to 2010 spanning equinox were derived from magnetic data by During the first two high-latitude intervals the two periods were well separated at~10.6 hr for the northern PPO system and~10.8 hr for the southern. During the third interval studied here the periods were much closer at~10.64 hr for the northern and~10.69 hr for the southern, with the northern period still shorter than the southern. Directly after the end of the this interval, however, in mid-2013 the two periods coalesced for~1 year and then reversed sense with the northern period longer than the southern after mid-2014. Figure 3c shows the beat period, τ B , of the two PPO oscillations plotted on a log scale, given by For the first two high-latitude intervals (2006/2007 and 2008) the beat period was relatively short at~20 days, compared with an orbit period of~10 days (Figure 2d), such that successive passes typically occurred at The north/south amplitude ratio, k, of the two PPOs determined from beat phase modulation of the oscillation phases within the equatorial region when available, where k is plotted directly between 0 and 1 in the lower half plot and 1/k is plotted between 1 and 0 in the upper half plot (see also the k scale on the right). Dotted lines indicate intervals in which southern PPO modulations could not be discerned, indicating k ≥ 5. (e) The amplitudes (nT) of the northern (blue) and southern (red) oscillations in the equatorial field region, specifically for the azimuthal field component. Where the two amplitudes are equal (i.e., k = 1), they are indicated by red and blue dashed lines.
intervals of a half of a beat period. For the third high-latitude interval (2012/2013), however, due to the smaller separation of the periods the beat period increased to~100 days compared with the orbit period of~10-20 days, such that successive passes occurred at intervals of a tenth to a fifth of a beat period. Even so, approximately four beat cycles were completed over the whole interval of study, such that the full range of relative phase conditions between the two PPO systems was experienced. Figure 3d shows the north/south amplitude ratio, k, of the two PPO oscillations determined from beat phase modulation of the oscillation phases within the quasi-uniform perturbation field region where the two PPO oscillations are superposed (e.g., Andrews et al., 2012;Provan et al., 2013Provan et al., , 2014Provan et al., , 2016. (Observations in the quasi-dipolar regimes cannot be used for this purpose since no model of the polar amplitude variation, e.g., with radial distance, has yet been devised.) Parameter k is plotted directly in the range 0 ≤ k ≤ 1 in the lower half of the plot, while 1/k is plotted in the range 1 ≥ /k ≥ 0 in the upper half, thus uniformly covering the full range of k from 0 to infinity. Dotted lines indicate intervals in which southern PPO modulations could not be discerned in these data, indicating k ≥ 5 (Provan et al., 2013). During southern summer the southern PPO was dominant by a factor of~2.5, with the two oscillations becoming near equal in amplitude in the trans-equinoctial interval between 2008 and 2011 (Andrews et al., 2012).
Postequinox, the values become more variable over~100 to 200 day intervals of southern dominant, northern dominant, and near-equal amplitude oscillations, before becoming clearly northern dominant soon after the start of 2015 during the interval of reversed periods. In the second and third high-latitude intervals of primary interest here the oscillations were of near-equal amplitude in both cases, being slightly southern dominant with k ≈ 0.87 in 2008 and slightly northern dominant with k ≈ 1.6 falling to k ≈ 1.0 in 2012/2013. These values were both determined from beat modulation of unfiltered azimuthal field data obtained during the rapid near-periapsis equatorial passes on these highly inclined Revs (Andrews et al., 2012;Provan et al., 2016). Individual amplitudes of the equatorial azimuthal field oscillations are plotted in Figure 3e, showing that these decreased between the second and third high-latitude intervals, by a factor of~0.7 for the northern system and~0.6 for the southern system. The PPO-related currents may also be expected to be less by similar factors.

Data Set
As discussed in section 1, the data set analyzed here consists of 1-min averaged values of the azimuthal magnetic field, B φ , obtained on the periapsis passes of Cassini Revs 168-194 between mid-2012 and mid-2013, corresponding to the central interval of Saturn northern spring. Since data from Rev 172 are missing due to an extended trans-periapsis data gap, this gives a set of 26 passes across both northern and southern field-aligned current regions. However, as in previous studies of the 2008 data by Hunt et al. (2014Hunt et al. ( , 2015, initial examination shows that the northern passes on Revs 173 and 182 and the southern pass on Rev 169 show atypical signatures, such that these have been excluded from further analysis. The data from the three excluded passes are briefly examined in Appendix A and compared with the included data in Figure A1. They are taken to represent cases of the occurrence of relatively short-lived (one~1 day pass or less) unusual dynamic activity taking place in the magnetosphere on these passes, such as the effects of auroral storms associated with strong magnetospheric compressions by the solar wind (e.g., Bunce et al., 2010). Overall, the data set employed here thus consists of 24 northern and 25 southern polar passes. This compares with 34 northern and 31 southern passes in the 2008 data set (Hunt et al., 2014(Hunt et al., , 2015.  Figure 4d is through the planet from the north. The break point between the two mappings lies at the near-periapsis point where the trajectory is tangent to the model flux shells, such that the mapped footprints then lie at the largest colatitudes from the poles in both hemispheres. As in Hunt et al. (2014Hunt et al. ( , 2015, the model field employed is axisymmetric, consisting of the three-term internal planetary field model of Burton et al. (2010), together with the empirical ring current model of Bunce et al. (2007) for a typical subsolar magnetopause distance of 22 R S . The ionosphere is taken to lie 1,000 km above the International Astronomical Union (IAU) 1-bar Saturn reference spheroid (e.g., Galand et al., 2011), the polar and equatorial radii of the 1-bar spheroid being 54,364 and 60,268 km, respectively. Also shown are the average locations of the auroral regions both north and south, as determined from Cassini Ultraviolet Imaging Spectrograph data by Carbary (2012), where the central line with black circles at each LT hour shows the colatitude of maximum intensity, while the lines on either side show the half-maximum intensity points. Figure 4d shows that inbound the spacecraft crossed southern auroral field lines centered near~17.5°southern colatitude (where the main upward currents flow) in the premidnight and midnight sectors. Figure 4e shows that outbound it crossed northern auroral field lines at~16°n orthern colatitude in the postmidnight and dawn sectors. The difference in colatitude between the two hemispheres is due to the axial quadrupole asymmetry of the internal field.   where Z is again distance along Saturn's spin/magnetic axis and ρ is the perpendicular distance from that axis. Arrowed black lines show the field lines of the magnetic model, plotted at 5°ionospheric colatitude intervals from the southern pole to 30°(~3.6 R S in the equatorial plane). The colored areas, also bounded by field lines, indicate the approximate locations of principal field-aligned current regimes, specifically those of the southern subcorotation system found by Hunt et al. (2014). Pale blue areas show the polar downward current regions (Figure 1b), the darker blue bands the region of enhanced downward current mapping from near the OCB to~15.5°and 17°in the northern and southern ionospheres respectively, and the red shaded area the main upward current region mapping to~17°and 19°in the northern and southern ionospheres, respectively (extending in to~10 R S in the equatorial plane). The main PPO currents also flow in the darker blue and red regions (Figures 1f and 1i).  Hunt et al. (2014Hunt et al. ( , 2015, this is primarily defined by the trans-periapsis segment lying between the two points with the smallest mapped colatitudes in the two hemispheres, that is, the points where the trajectory in the ρ-Z plane is tangent to the polar field lines. For the 2012/2013 data set we have also applied a limiting radial distance of 15 R S , corresponding approximately to the outer limit of where the model field, dominated by the internal field, is likely to be reasonably accurate, representing a radius roughly half way to the magnetopause in the dayside magnetosphere. This limiting distance was not exceeded on any pass in the 2008 data set but applies to a few of the earlier Revs in the 2012/2013 interval as seen in Figure 5a. Figure 5a shows that the 2012/2013 Revs provide relatively straightforward cuts through the main fieldaligned current regions, as for the 2008 data set, except perhaps for the early Revs in the northern hemisphere. However, the larger periapsis distances for the 2012/2013 data set imply slower crossings than for the 2008 data set, the consequences of which will be shown in section 2.3. With crossings further from the ionosphere, the ionospheric currents deduced from the field data may also be overestimated or underestimated if there are significant cross-field currents flowing between the ionosphere and the point of observation (Figures 1b, 1e, and 1h). However, much of the cross-field current is expected to flow closer to the equator where the bulk of the plasma mass is located, far from the field-aligned current crossing points. In particular, in the equatorial regions of these Revs at radial distances of~6-9 R S , most of the plasma mass (and hence any cross-field current) is expected to be confined to within~1 R S of the equatorial plane (Persoon et al., , 2013.

Data Analysis Procedures
Analysis of the B φ data from 2012/2013 follows that in the related studies by Hunt et al. (2014Hunt et al. ( , 2015, such that only a brief outline will be given. First, we estimate and remove the generally small contributions to B φ that are not due to field-aligned currents. Since Saturn's internal and inner ring current fields are closely axisymmetric (Burton et al., 2010;Kellett et al., 2011), these fields are expected to be due principally to the tail and magnetopause currents. As in previous studies, these have been estimated on the Cassini trajectory using the paraboloid model of Saturn's magnetosphere (see, e.g., Alexeev et al., 2006), specifically using the "intermediate" model of Belenkaya et al. (2008) for a subsolar magnetopause distance of 22 R S . These fields have then been subtracted from the observed values, such that all subsequent analysis uses the "corrected" values. However, the subtracted fields are generally small compared with the observed field (section 2.3), such that this procedure has only a modest effect on the results.
The corrected azimuthal field is then mapped along field lines into the appropriate ionosphere and the meridional current at the feet of the field lines estimated (Figure 1). To do this, we assume that ρB φ is constant along field lines (ρ again being perpendicular distance from the spin/magnetic axis), an approximation that is exact for axisymmetry and in the absence of cross-field current flow on the field line extending between the observation point and the ionosphere. Hunt et al. (2014) showed that the approximation should also be valid for sheet-like currents such as those considered here, where the latitude scale is much smaller than the longitude scale. The azimuthal field just above the ionosphere, B φi , is then given by where ρ i is the perpendicular distance of the foot of the field line in the ionosphere. Ampère's law then gives the meridional (mainly Pedersen) current per radian of azimuth in the ionosphere at the feet of the field lines which switches off the azimuthal field under the ionosphere, I m , as where μ 0 is the permeability of free space. The upper and lower signs correspond to the northern and southern hemispheres, respectively, and I m is taken positive when directed toward the equator in both hemispheres. Again, assuming that longitude variations can be neglected, current continuity requires that field-aligned currents flow into and out of the ionosphere as I m varies with colatitude. For meridional currents at colatitudes θ 1 and θ 2 with respect to either pole, θ 2 > θ 1 , the amount of field-aligned current flowing between the two locations, I ‖ , is given by where negative values indicate downward currents in both hemispheres and positive values upward currents. Thus, if I m (θ 2 ) < I m (θ 1 ), net field-aligned current flows out of the ionosphere between these two points, while if I m (θ 2 ) > I m (θ 1 ), net field-aligned current flows into the ionosphere between them. The field-

10.1029/2017JA024885
Journal of Geophysical Research: Space Physics aligned current then remains fixed on a given flux tube (such that (j ‖ /B) = constant, where j ‖ is the current density) in the region between the ionosphere and the magnetospheric closure currents.

Representative Examples
The procedures in section 2.2 result in the estimation of colatitude profiles of B φi (and hence I m ) in both northern and southern hemispheres on each pass. Figure 6 shows examples for the representative Revs in Figure 4. Figures 6a and 6b Jinks et al. (2014), at 13.3°colatitude in the northern hemisphere and 15.6°in the southern hemisphere. In the upper panels the black lines show the azimuthal field as mapped to the ionosphere using equation (2), while the colored lines, blue in the north and red in the south, show the profiles after removal of the estimated magnetospheric contribution, as discussed in section 2.2. The effect is seen generally to be small, though the differences increase with distance away from the noon-midnight meridian.
The lower panels in Figures 6a-6f show the northern and southern PPO phases, Ψ N (blue) and Ψ S (red) as determined from analysis of the PPO perturbation fields observed in the near-equatorial and polar regions by Provan et al. (2013Provan et al. ( , 2016. The analysis yields a model of Φ N,S (t), the azimuthal angle of the Ψ N,S = 0°F  (2) and plotted versus northern ionospheric colatitude θ iN , while in the upper panels of Figures 6b, 6d, and 6f data from south of the planetary equator are similarly projected into the southern ionosphere and plotted versus southern ionospheric colatitude θ iS . Black data show the fields derived directly from the azimuthal field as measured, while the blue (northern) and red (southern) data show the fields after subtraction of the magnetospheric contribution estimated using the paraboloid field model, as discussed in the text. Vertical dashed lines indicate the position of the northern and southern OCB . The lower parts of the figures show the corresponding phases of the northern (blue) and southern (red) PPO systems at the spacecraft, Ψ N,S modulo 360°, given by equation (5). meridians (see Figures 1d and 1g) measured from noon in the sense of planetary rotation. The northern and southern phases in Figure 6 are then given by where φ is the azimuth of the spacecraft relative to noon, also measured in the sense of planetary rotation. The PPO phases are seen to vary significantly on a given pass, related to significant variations of the observed azimuthal field equatorward of the OCB, such that as indicated in section 2.1, individual profiles do not approximate to snapshots of the field at a given time. As in the study by Bunce et al. (2014), the varying PPO phase on a given pass must thus be taken into account. These phase plots also illustrate the varying PPO beat phase conditions encountered, with near antiphase for Rev 170, near in phase for Rev 180, and near quadrature for Rev 190.   (2)) and I m (equation (3)) data, respectively, plotted versus northern colatitude θ iN , while Figures 7c and 7d show the southern hemisphere B φi and I m data, respectively, plotted versus southern colatitude θ iS . The averaged position of the OCB in each hemisphere is shown Overall, the azimuthal fields are seen generally to be negative in the north and positive in the south, indicative of lagging fields and subcorotation flows, though with major variability in the auroral region equatorward of the OCB leading to intervals with leading fields associated with supercorotating flows with respect to the planet's upper atmosphere. As discussed further below, the mean values of these data shown by the black lines provide an initial estimate of the subcorotation-related (PPO-independent) fields and currents in each hemisphere. The variations about the means also provide an indication of the amplitude of the PPO effects, which are seen to be comparable in magnitude. Other sources of variability are no doubt present as well, particularly at smaller spatial scales, but have been mitigated by the exclusion of atypical field/current profiles as discussed in section 2.1 and Appendix A.
If the PPO-related fields vary approximately sinusoidally with the PPO phases Ψ N,S as discussed in section 1.1, they will cancel near to zero at a particular colatitude given good phase coverage in the overall data set. As indicated above, the mean values of the data shown by the black lines in Figure 7 thus provide an initial estimate of the subcorotation-related current profiles (as demonstrated in section 3.4). These means have been calculated in overlapping 1°colatitude bins at intervals of 0.5°, where the data from each contributing Rev were first averaged individually, and then the means taken of these Revaveraged values. A mean value is not shown if a bin contains data from less than three Revs. The error bars show the standard error of the means of the contributing Rev-averaged values. Examining these mean values, it is seen that the rising values of the current with colatitude in both polar regions, peaking at~1 MA rad À1 just equatorward of the OCB, implies the existence of downward directed field-aligned currents in these regions totaling~6 MA in each hemisphere assuming axisymmetry. The same current then flows out of the ionosphere over the auroral regions equatorward of the OCB, with the meridional current reducing to near-zero mean values beyond~18°in the northern hemisphere and~20°in the

10.1029/2017JA024885
Journal of Geophysical Research: Space Physics southern. In the two polar regions the current profiles are seen to be different, however, indicative of a broadly distributed downward current in the northern (summer) hemisphere, while the meridional currents remain small to within a few degrees of the OCB in the southern (winter) hemisphere, indicative of weak ionospheric conductivities, and then rise to similar values as in the northern auroral region due to a narrower ring of downward current. A similar polar region asymmetry was found by Hunt et al. (2015) in the 2008 data during southern summer conditions but reversed in sense between the two hemispheres (their Figure 3). These evidently seasonal asymmetries will be discussed further in sections 3.8 and 4.

PPO-Related Variations at Small and Large Colatitudes
We now examine the data in more detail for PPO-related variations, beginning by examining B φi at small and large colatitudes, away from the main current sheets, where the mean values in Figure 7 do not vary greatly with colatitude. For the northern hemisphere we employ polar data from θ iN ≤ 12°and nearer-equatorial data from θ iN ≥ 17.5°, while for the southern hemisphere we employ polar data from θ iS ≤ 13°and nearerequatorial data from θ iS ≥ 19°.  Hunt et al. (2014Hunt et al. ( , 2015 found dual modulation in the northern hemisphere data but not in the southern.
As discussed in section 1.1, we expect the PPO azimuthal field to vary as + sin Ψ N,S in the equatorial regions and as À sin Ψ N,S on the opposite side of the field-aligned currents in the corresponding northern or southern polar regions. The black lines in Figure 8 show best fit sinusoids given by where the fit parameters are the mean value 〈B φi 〉 (black dashed line), amplitude B φi0 , and phase Δ. The fits were obtained by Levenberg-Marquardt least squares minimization (Markwardt, 2009), with fit parameters and root-mean-square (RMS) deviations being given in the upper left of each figure. Uncertainties were estimated using the "jackknife" statistical resampling method (section 4.1 of Hunt et al., 2015), where the circular standard deviation is given for phase Δ (Mardia & Jupp, 2000).
Examination of the polar data in Figures 8a and 8e shows PPO modulations close to the expected À sin Ψ N forms in the northern hemisphere and À sin Ψ N,S in the southern hemisphere (Δ close to 180°in both cases), though the northern modulations are weak compared, for example, with the~30 nT amplitude of the corresponding southern polar oscillations in the 2008 data ( Figure 5a of Hunt et al., 2015). Similar weak oscillations are derived when the northern polar data are plotted versus southern phase (Figure 8c) and the southern polar data versus northern PPO phase (Figure 8g), with phases Δ that do not correspond to simple physical expectation. Previous studies of the 2008 polar data by Andrews et al. (2012) and Hunt et al. (2015) did not detect modulation from the opposite hemisphere within a 10% limit by amplitude. For the near-equatorial data shown in Figures 8b, 8d, 8f, and 8h, however, clear dual modulation close to the expected + sin Ψ N,S form is seen in both hemispheres (Δ close to 0°modulo 360°in all cases), with the northern PPO modulation being stronger than the southern in the northern hemisphere, while the two modulations are near equal in amplitude in the southern hemisphere. However, if significant cross-field currents flow between the spacecraft and the corresponding magnetically conjugate ionosphere, particularly in the southern hemisphere due to the larger distances at which the auroral field lines are crossed (Figure 5a), the amplitudes in Figures 8b and 8f driven from the corresponding hemisphere could be underestimated, while those in Figures 8d and 8h driven from the opposite hemisphere could be overestimated. Nevertheless, these results show the presence of dual modulation on closed field lines in both hemispheres (see Figure 1), which must be taken into account in further analysis. Figure 8 also shows that there are no clear phase offsets between the equatorial and polar modulations in these data, unlike the~35°offsets found in the 2008 data (Hunt et al., 2015).

10.1029/2017JA024885
Journal of Geophysical Research: Space Physics

Division of Data Into PPO "Phase Squares"
To determine the PPO dependency of the dual-modulated currents, we follow the procedure adopted by Hunt et al. (2015) to analyze the dual-modulated 2008 northern hemisphere currents but applied here to the data from both hemispheres. The data from each hemisphere are thus divided into non-overlapping 90°wide bins of northern and southern PPO phase, centered on Ψ N,S = 0°, 90°, 180°, and 270°(modulo 360°), each bin thus being ±45°about these values. We refer to these bins as phase squares, there being 16 in total for each hemisphere, corresponding to four northern phase bins times four southern phase bins. We then note that since the PPO-related modulations vary approximately as ± sin Ψ N,S (Figure 8), northern system modulations should average near to zero in phase squares centered on Ψ N = 0°and 180°, while southern system modulations should average near to zero in phase squares centered on Ψ S = 0°and 180°. On the other hand, since the mean value of ± sin Ψ N,S in the ±45°ranges about 90°and 270°is ± sinΨ N;S ¼ ±2 ffiffi ffi 2 p =π≈±0:90, northern system modulations should average close to the maximum and minimum values (actually~90% thereof) in the phase squares centered on Ψ N = 90°and 270°, while southern system modulations should average close to the maximum and minimum values in the phase squares centered on Ψ S = 90°and 270°. Of course, in practice, these averages will depend in detail on the actual distribution of PPO phases within the data set, which Figures 7 and 8 indicate are in general broadly spread as required.

Northern Hemisphere Data
In Figure 9 we thus show northern colatitude profiles of the northern hemisphere meridional current in a similar format to Figure 7b but now separated into the 16 phase squares described above, arranged in four rows of four panels that display the effects of individual current components as outlined below. Due to the relatively slow traversals of the polar field lines, these profiles typically consist of short data segments from multiple Revs (color coded as in Figures 7 and 8). Despite this and the consequent data scatter, the data set is evidently sufficient to approximately define most of the current profiles in each phase square. As in Figure 7, mean values are computed in overlapping 1°colatitude bins every 0.5°(black lines and circles) by first averaging the data from each contributing Rev in each bin and then taking the mean of these Revaveraged values. Means are plotted in the figure when there are at least two contributing Revs in a bin.  Figures 9i-9l (third row) show data from the four phase squares for which Ψ N = 0°or 180°such that the effects of the northern PPO system should average near to zero, but where Ψ S = 90° (Figures 9i and 9j) or 270° (Figures 9k and 9l), which should show equal and opposite effects of the southern PPO currents superposed on the subcorotation current. Again, it is seen that the currents for Ψ S = 90°in Figures 9i and 9j extend to negative values at larger colatitudes while remaining positive for Ψ S = 270°in Figures 9k and 9l, thus exhibiting the effect of the southern PPO current system expected from Figure 1f. These profiles thus confirm the dual modulation of the northern hemisphere currents by the two PPO systems. Figures 9m-9p (fourth row) show the final set of four phase squares for which both Ψ N = 90°o r 270°and Ψ S = 90°or 270°, such that near maximum/minimum effects of both PPO systems should simultaneously be present, either in phase with each other (Figures 9m and 9n), then showing similar PPO-related effects to those just described, or in antiphase (Figures 9o and 9p) where they part cancel. Figure 10 similarly shows southern colatitude profiles of the southern hemisphere meridional current divided into 16 nonoverlapping phase squares, as in Figure 9.  (Figures 9e and 9f) and Ψ N = 270°with Ψ S = 0°or 180° (Figures 9g and 9h). (i-l) Phase squares for which the southern but not the northern PPO currents are expected to contribute, Ψ S = 90°with Ψ N = 0°or 180° (Figures 9i and 9j) and Ψ S = 270°with Ψ N = 0°or 180° (Figures 9k and 9l). (m-p) Phase squares for which both the northern and the southern PPO currents are expected to contribute, either in phase with each other Ψ N,S = 90°or 270° (Figures 9m and 9n) or antiphase (Figures 9o and 9p).  ) data, derived from southern hemisphere mapped azimuthal magnetic field data (equation (3)), are shown separated into 16 nonoverlapping phase squares of northern and southern PPO phases Ψ N and Ψ S (deg) and plotted versus southern ionospheric colatitude θ iS (deg). The format is the same as Figure 9.

10.1029/2017JA024885
Journal of Geophysical Research: Space Physics and become small or negative in Figures 10g, 10h, 10k, and 10l, in line with expectations from Figure 1i. Figures 10m-10p (fourth row) show the cases where both PPO systems are simultaneously present, either in phase with similar effects being evident or antiphase. These figures again demonstrate dual PPO modulation of the southern currents, as in Figure 8.

Northern and Southern Separated Subcorotation and PPO-Related Current Profiles 3.4.1. Subcorotation Current Profiles
The data in Figures 9 and 10 can be used to estimate separated colatitude profiles of the subcorotation and PPO-related currents in the northern and southern hemispheres. Subcorotation currents may first be estimated by combining data from the phase squares for which the PPO effects should be small and average near to zero, that is, Figures 9a-9d and 10a-10d for the northern and southern hemispheres, respectively. Profiles of mean values for the northern and southern data are shown in Figures 11a and 12a, respectively, together with uncertainty estimates given by the standard error of the means. These were computed in the same manner as for Figure 7, with mean values being shown provided at least three Revs contribute to a bin. As may be expected, these profiles are similar to the overall mean profiles in Figures 7b and 7d, respectively. In the northern (summer) hemisphere in Figure 11a a broadly distributed downward current totaling~6.0 MA (assuming axisymmetry) flows in the polar region to~15°northern colatitude,~1.5°equatorward of the OCB, which returns up the field lines between~15°and~17.5°centered near the auroral oval ( Figure 4e). In the southern (winter) hemisphere in Figure 12a a similar total current flows downward in a narrower polar ring between~11°and~15.5°near the OCB, which largely returns up the field lines between~17°a nd~20°. Further discussion is given in sections 3.6 and 3.8.

PPO Current Profiles
An estimate of the northern PPO current profile in the northern ionosphere can also be obtained by combining the data in Figures 9e and 9f, for which Ψ N = 90°with Ψ S = 0°and 180°, and separately the data in Figures 9g and 9h, for which Ψ N = 270°with Ψ S = 0°and 180°, then subtracting these combined profiles and dividing by two, thus eliminating the common northern subcorotation current. The mean profiles and uncertainties in each case were again derived as for Figure 7, with mean values being retained provided there were at least three contributing Revs. Specifically, the Ψ N = 270°profile is subtracted from the Ψ N = 90°profile to obtain the northern PPO profile for Ψ N = 90°shown by the blue dotted line in Figure 11b and vice versa to obtain the equal but opposite profile for Ψ N = 270°shown by the red dotted line in Figure 11b. Similarly, the southern PPO current profile in the southern ionosphere can be obtained by combining the data in Figures 10i and 10j for which Ψ S = 90°with Ψ N = 0°and 180°, and the data in Figures 10k and 10l for which Ψ S = 270°with Ψ N = 0°and 180°, and subtracting and dividing by 2 to yield the red and blue dotted line profiles in Figure 12c. Both northern and southern hemisphere PPO profiles indicate modest near-uniform currents~0.2 MA rad À1 at polar latitudes, which reverse in sense to peak 4°-5°equatorward of the OCB at~1 MA rad À1 and then modestly decline at larger colatitudes. However, in both cases the profiles show unexpected intervals of near-zero current, equatorward of the OCB for the northern PPO profile in Figure 11b and poleward of the OCB for the southern profile Figure 12c, the possible origin of which is discussed below.
The southern PPO current in the northern ionosphere can be estimated by combining the data in Figures 9i and 9j for which Ψ S = 90°with Ψ N = 0°and 180°, also combining the data from Figures 9k and 9l for which Ψ S = 270°with Ψ N = 0°and 180°, and subtracting and dividing by 2, with results shown by the blue (Ψ S = 90°) and red (Ψ S = 270°) lines in Figure 11c. Similarly, the northern PPO current in the southern ionosphere can be estimated by combining the data from Figures 10e and 10f for Ψ S = 90°with Ψ S = 0°and 180°, also combining the data from Figures 10g and 10h for Ψ N = 270°with Ψ S = 0°and 180°, and subtracting and dividing by two, with results shown by the blue (Ψ N = 90°) and red (Ψ N = 270°) lines in Figure 12b. The resulting profiles in both cases show variable near-zero currents in the polar regions, rising to peak~3°-4°e quatorward of the OCB at~0.6-0.7 MA rad À1 , before decreasing more slowly again at larger colatitudes.
We briefly note that the data in Figures 9e-9l and 10e-10l can also be employed to generate additional independent estimates of the subcorotation currents, if instead of subtracting the combined profiles for opposite PPO phases they are added and divided by 2, such that the expected oppositely signed PPO effects then near cancel. Examination shows that these estimates are indeed very similar to those in Figures 11a and 12a, though with larger uncertainties, such that they are not shown here.

10.1029/2017JA024885
Journal of Geophysical Research: Space Physics

.3. Correction of PPO Current Profiles for the Effects of Latitudinal Oscillation
The results in Figures 11b (dotted lines) and 12b for the northern PPO currents in the northern and southern ionospheres, respectively, and 12c (dotted lines) and 11c for the southern PPO current in the two hemispheres, both appear to suggest rather different current profiles in the two hemispheres. For the northern PPO system, for example, the current profile deduced in the northern hemisphere has near-zero values within several degrees of the OCB, while in the southern hemisphere the current grows gradually away from the OCB. Similarly, the southern PPO current deduced in the southern hemisphere reverses sense~2.5°equatorward of the OCB, while the southern PPO current in the northern hemisphere again grows gradually away from the OCB. A possible contributor to these differences lies in an effect observed in the well-characterized 2008 southern hemisphere data set investigated by Hunt et al. (2014), for which dual modulation was not significant. This study revealed that the current layers in the southern hemisphere oscillated in latitude near sinusoidally, being displaced by~1°poleward near Ψ S ≈ 90°and~1°e quatorward near Ψ S ≈ 270°, an effect attributed to the action of the rotating twin-vortex flow in the southern hemisphere shown by the red dashed lines in Figure 1g. Such an effect is far more difficult to observe directly in the data employed here for reasons discussed in section 2. Given the expected flows in the northern ionosphere shown by the red dashed lines in Figure 1d, the anticipated equivalent effect in the northern hemisphere is a comparable equatorward displacement of the current layers at Ψ N ≈ 90°and a poleward displacement at Ψ N ≈ 270°.
If we consider the effect of such shifts on the differenced profiles employed to estimate the northern PPO current in the northern ionosphere (dotted lines in Figure 11b) and the southern PPO profile in the southern hemisphere (dotted lines in Figure 12c), it is readily seen that the effect on the oppositely signed PPO currents themselves is small, serving only to modestly broaden the current profile. However, similar shifts of the subcorotation current profile, followed by subtraction and division by 2, result in spurious contributions to the PPO-dependent current profiles. Specifically, in the northern hemisphere it leads for a phase of Ψ N ≈ 90°to the addition of a negative current in the poleward region where the subcorotation current increases with colatitude and a positive current in the equatorward region where the subcorotation current decreases with colatitude, with equal and opposite effects for Ψ N ≈ 270°. In the southern hemisphere the senses of these effects are reversed. Addition of such spurious differences to the expected PPO currents then reduces the expected PPO-related currents in both hemispheres, thus potentially explaining the suppression of the PPO currents in Figures 11b and 12c, particularly in those colatitude intervals where the subcorotation current varies strongly with colatitude (see Figures 11a and 12a). This effect does not apply to the southern PPO profile in the northern ionosphere (Figure 11c), however, nor the northern PPO profile in the southern ionosphere (Figure 12b), since these were determined for Ψ N = 0°and 180°in the northern hemisphere and for Ψ S = 0°and 180°in the southern hemisphere, for which the displacements are expected to be near zero. This assumes, of course, that the northern PPO system flow driven in the southern ionosphere along field lines from the north, and vice versa, are secondary effects.
To investigate these effects, we have appropriately shifted the profiles in colatitude by 1°to bring them into alignment for Ψ N,S = 90°and 270°before subtraction and division by two. Specifically, in the northern hemisphere we have shifted the combined Ψ N = 90°profile poleward by 1°and the combined Ψ N,S = 270°profile equatorward by 1°before subtraction, while in the southern hemisphere we have shifted the combined Ψ S = 90°profile equatorward by 1°and the combined Ψ S = 270°profile poleward by 1°before subtraction. The modified corrected current profiles are shown by the solid blue and red lines in Figures 11b and 12c. In the northern hemisphere the overall profiles now reverse sign directly from small polar values to much larger auroral region values~2°equatorward of the OCB. Similarly in the southern hemisphere, the unexpected interval of weak polar values just poleward of the OCB is removed, the polar currents now being essentially continuous, while weaker currents occur in the reversal region~2°equatorward of the OCB. Overall, these corrected profiles are in better agreement with each other and with expectation, and will hence be adopted here in further discussion. Figure 11d shows our estimate of the overall northern hemisphere profile at the peaks of the northern PPO current but in the absence of the southern PPO current, while Figure 11e correspondingly shows the overall northern hemisphere profile at the peaks of the southern PPO current in the absence of the northern PPO current. That is, Figure 11d corresponds to Ψ N = 90°(blue) and Ψ N = 270°(red) with Ψ S = 0°and 180°,

10.1029/2017JA024885
Journal of Geophysical Research: Space Physics derived by combining the profile for the subcorotation current in Figure 11a, reproduced in black in Figures 11d and 11e, with the corresponding corrected profiles for the PPO-related currents in Figures 11b  and 11c. In Figure 11d we note that the addition of the northern PPO current nearly doubles the upward directed current in the auroral region when Ψ N = 90°, while considerably reducing it over the colatitude range shown when Ψ N = 270°. Figure 11e shows that the southern PPO currents in the northern hemisphere have a similar effect but of lesser magnitude. Figures 12d and 12e show corresponding results in the southern hemisphere for the overall current at the peaks of the northern PPO current in the absence of the southern PPO current and at the peaks of the southern PPO current in the absence of the northern PPO current, respectively. Again, major modulations of the current profiles equatorward of the OCB are observed, with stronger upward currents for Ψ N,S = 270°and weaker for Ψ N = 90°.
We finally recall that due to the finite width of the phase square bins the PPO-related profiles in Figures 11  and 12 will generally represent values slightly smaller than those at the extrema at Ψ N,S = 90°and 270°. If the variations are near sinusoidal in PPO phase, however, the expected factor is only 2 ffiffi ffi 2 p =π≈0:90, as mentioned in section 3.3. The PPO values in these figures can be multiplied by π=2 ffiffi ffi 2 p ≈1:11 for a better estimate of the true peak if required.

Comparison of Model Current Profiles With Dual-Modulated Data
The derivation of the northern and southern hemisphere current profiles in Figures 11 and 12 employed data from 12 out of the total of 16 phase squares in Figures 9 and 10. The subcorotation profiles were derived using the phase squares for which neither northern nor southern PPO system should on average be present, while the individual PPO-related profiles were derived from those in which one system should be present, but not the other. The data from the remaining 4 phase squares for each hemisphere that are expected to be dual modulated by both PPO systems (Figures 9m-9p and 10m-10p), either in phase or antiphase with each other, were not employed in the derivation of Figures 11 and 12, thus providing the opportunity for an independent check on the results. That is, to say, we can synthesize northern and southern hemisphere model current profiles corresponding to in-phase and antiphase dual-modulated PPO conditions from the results in Figures 11a-11c and 12a-12c and compare them with the independent northern and southern hemisphere data in Figures 9m-9p and 10m-10p.
Results are shown for the northern hemisphere in Figure 13, for in-phase PPO conditions in Figures 13a-13c and antiphase in Figures 13d-13f. Figure 13a shows the combined PPO-related current profiles for the cases where Ψ N and Ψ S are both 90°(blue) and both 270°(red), obtained from Figure 11b (corrected) profiles and 11c, together with the subcorotation current profile from Figure 11a (black). Figure 13b shows the combined subcorotation plus PPO current profiles with Ψ N and Ψ S both 90°, together with the color-coded data from Figure 9m, while Figure 13c shows the combined subcorotation plus PPO current profiles for Ψ N and Ψ S both 270°, together with the data from Figure 9n. In accordance with the discussion in section 3.4.3 the former profile has also been shifted 1°equatorward and the latter 1°poleward, though this does not strongly affect the comparison. Despite the data scatter, the overall agreement is seen to be reasonable, with slowly increasing positive polar values decreasing to negative auroral region values in Figure 13b, while smaller positive polar values increase to larger values in the auroral region before falling again in Figure 13c. Figure 13d similarly shows the combined PPO current profiles when the two systems are in antiphase, for the cases where Ψ N = 90°and Ψ S = 270°(blue line) and Ψ N = 270°and Ψ S = 90°(red line), again obtained from Figures 11b (corrected profile) and 11c, together with the subcorotation current from Figure 11a (black line). Figure 11e then shows the overall current profile corresponding to antiphase conditions Ψ N = 90°and Ψ S = 270°(again shifted 1°equatorward) together with the color-coded data from Figure 9o, while Figure 11f shows the overall profile for the opposite antiphase condition Ψ N = 270°and Ψ S = 90°(again shifted 1°poleward) together with the data from Figure 9p. The overall agreement is again seen to be reasonable, with a sharp switch from larger positive polar to near-zero values in Figure 11e and smaller currents with similar behavior occurring in Figure 11f. Figure 14 shows results for the southern hemisphere currents, derived from combinations of subcorotation and PPO-related current profiles shown in Figure 12 (the corrected profiles in Figure 12c), again shifted by 1°poleward for Ψ S = 90°and 1°equatorward for Ψ S = 270°, together with the color-coded data from Figures 10m-10p. Reasonable overall agreement is again noted, with a large near-monotonic increase in  ) derived from the results in Figure 12 for cases of dual modulation by both northern and southern PPO systems, with the corresponding independent data shown in Figures 10m-10p. The cases for which the two systems are in phase, Ψ N,S = 90°and 270°, are shown in Figures 14a-14c, while the cases for which the two systems are in antiphase, Ψ N = 90°w ith Ψ S = 270°and Ψ N = 270°with Ψ S = 90°are shown in Figures 14d-14f. The format is the same as for Figure 13.

10.1029/2017JA024885
Journal of Geophysical Research: Space Physics 3.6. Comparison of Current Profiles in the Northern and Southern Hemispheres Figure 15 compares the subcorotation and PPO-related current profiles in the two hemispheres. The northern hemisphere currents are shown by solid lines plotted versus northern ionospheric colatitude, while to allow detailed comparison the southern hemisphere currents have been mapped to the northern ionosphere to points of equal enclosed polar magnetic flux (equivalent to "along field lines" in the closed field region) and are shown by crosses and dotted lines. An equivalent (slightly nonlinear) scale of southern ionospheric colatitude is shown at the top of the figure. To avoid clutter, uncertainty values are shown only for the northern profiles, noting that the uncertainties for the southern profiles are entirely comparable (Figures 11 and 12). Vertical dashed lines show the northern (blue) and mapped southern (red) OCBs estimated by Jinks et al. (2014). In principle, these two boundaries must map to the same position as each other, enclosing the same amount of polar open flux, so that the difference seen in the figure represents a measure of their uncertainty. Figure 15a compares the northern and southern hemisphere subcorotation current profiles, whose form follows the previous discussion in section 3.1 (Figures 7b and 7d) and section 3.4 (Figures 11a and 12a). Here we see that the current values in the two hemispheres reach closely similar~1 MA rad À1 peak values on the same field lines in the vicinity of the OCB and just equatorward thereof but deviate from each other in opposite senses on either side of the boundary. In the southern polar region the current falls rapidly to small values 3°-4°poleward of the OCB while in the northern polar region the current falls much more gradually. In the equatorward region the current in the northern hemisphere falls rapidly to small values~2.5°equatorward of the peak, while in the southern hemisphere the current falls more gradually in this region. Figure 15b compares the northern PPO current in the "driving" northern hemisphere (solid lines) with the northern PPO current in the "responding" southern hemisphere (dotted lines). Similarly, Figure 15c compares the southern PPO current in the driving southern hemisphere (dotted lines) with the southern PPO current in the responding northern hemisphere (solid lines). In both cases, the near-constant polar values at smallest colatitudes in the driving hemispheres reverse sense across a fairly broad region equatorward of the OCB before peaking at~1 MA rad À1 in both hemispheres at larger colatitudes, and then falling again. Also in both cases, the current in the responding hemispheres increases more gradually with colatitude from near-zero values at and poleward of the OCB to peak at~0.7 MA rad À1 for the northern PPO system and~0.6 MA rad À1 for the southern PPO system and then fall again to values similar to those in the driving hemispheres. The difference between the driving and responding currents for a given system relates to the cross-field magnetospheric closure currents illustrated in Figures 1f and 1i, as will be discussed in section 3.7.
Figures 15d and 15e compare the driving and responding current profiles in the two hemispheres, respectively. Overall, the two systems have near-equal amplitudes in agreement with expectation from Figure 3c. The unexpected differences in detailed form between the driving and responding current profiles (see Figures 1f and 1i) may arise from averaging over a number of Revs with varying OCB positions, in the presence of a current reversal in the polar region in the case of the driving currents, but with near-zero current in the polar region in the case of the responding current.

Overall Current Continuity in the PPO Systems
We now give a brief discussion of current closure in the two PPO systems based on the results in Figure 15 and first consider continuity in the northern PPO system as indicated in Figure 15b. An overall schematic is provided in Figure 16a where the arrowed black lines show cross-field currents in both the ionosphere (short dashed circle) and magnetosphere, blue areas and arrows field-aligned currents flowing from the northern ionosphere to the southern, and red areas and arrows field-aligned currents flowing from the southern ionosphere to the northern. Arrows are labeled with the corresponding current in MA rad À1 as indicated below, referring specifically to peak values on the Ψ N = 90°À 270°meridian (Figure 1e). From Figure 15b it is seen that the main northern PPO field-aligned current in the northern hemisphere, flowing in a~3°colatitude band from just equatorward of the OCB to~17°(mapping to~11 R S in the equatorial plane) totals 1.2 MA rad À1 , directed upward at Ψ N = 90°and downward at Ψ N = 270°as indicated in Figure 16a. The contributions to this total current are~0.2 MA rad À1 from the polar ionosphere and~1.0 MA rad À1 from closed field lines at larger colatitudes. Of the total field-aligned current,~0.7 MA rad À1 , that is, somewhat over half, flows on these field lines into the southern ionosphere, downward at Ψ N = 90°and upward at Ψ N = 270°as  Jinks et al. (2014), the difference between them giving a measure of their uncertainty. For the PPO current profiles, those for Ψ N,S = 90°are shown blue, while the equal and opposite profiles for Ψ N,S = 270°are shown red. For clarity, standard error of the mean uncertainty estimates are shown only for the northern profiles, the comparable uncertainties for the southern profiles are shown in Figure 12. (a) Comparison between northern and southern hemisphere subcorotation (PPO-independent) current profiles (Figures 11a and 12a). (b) Comparison between northern and southern hemisphere northern PPO system current profiles (Figures 11b and 12b). (c) Comparison between northern and southern hemisphere southern PPO system current profiles (Figures 11c and 12c). (d) Comparison between the northern hemisphere northern PPO current profile with the southern hemisphere southern PPO current profile (Figures 11b and 12c). (e) Comparison between the northern hemisphere southern PPO current profile with the southern hemisphere northern PPO current profile (Figures 11c and 12b). NH = northern hemisphere; SH = southern hemisphere.
shown in the figure. The difference between these currents indicates that~0.5 MA rad À1 closes cross field in the magnetosphere in the region between. No significant additional contribution to the northern system current comes from the southern polar ionosphere. Equatorward of this main current layer, between~17°a nd~18.5°northern colatitude, corresponding to~19°and~20.5°southern colatitude (mapping betweeñ 11 and~8 R S in the equatorial plane), a small reversed "return" field-aligned current of~0.2 MA rad À1 flows in the southern ionosphere, upward at Ψ N = 90°, and downward at Ψ N = 270°, while on the same field lines in the northern ionosphere a larger return field-aligned current of~0.5 MA rad À1 flows, downward at Ψ N = 90°and upward at Ψ N = 270°. This difference indicates that of the total~0.5 MA rad À1 cross-field closure current estimated above,~0.3 MA rad À1 closes via return currents to the northern ionosphere (Figure 1e), shown for convenience of drawing as two contributions (black arrows) of 0.15 MA rad À1 either side of the equator in Figure 16a. The remaining~0.2 MA rad À1 , equal and opposite to the polar current, closes cross field in the magnetosphere to the opposite magnetic meridian, shown in the figure by the equatorial long dashed black arrows either side of the planet, indicating current flow out of and/or into the plane of the sketch. An ionospheric meridional current of~0.5 MA rad À1 remains flowing in both hemispheres at the large colatitude limit of the data at 20°northern and 22°southern colatitudes (~6.5 R S in the equatorial plane), indicated by the inner dotted field lines in Figure 16a. These lower- Figure 16. Sketches illustrating the overall current flow in the two PPO systems based on the results in Figures 15b and 15c. The arrowed black lines show cross-field currents in both the ionosphere (short dashed circle) and magnetosphere, blue areas and arrows field-aligned current flows from the northern ionosphere to the southern, and red areas and arrows field-aligned current flows from the southern ionosphere to the northern. Arrows are labeled with the corresponding current in MA rad À1 , referring specifically to maximum values on the Ψ N = 90°À 270°meridian for the northern system in Figure 16a (see Figure 1e), and on the Ψ S = 90°À 270°meridian for the southern system in Figure 16b (see Figure 1h), though the values given may be multiplied by a factor of~1.1 to account for averaging the data over phase squares. The numbers may also be converted to total azimuth-integrated currents of a given sense flowing either side of the planet by multiplying by a factor of~2.2 (section 3.7). The equatorial long dashed black arrows either side of the planet indicate the total cross-field current flowing in the magnetosphere between the two meridians into and out of the plane of the sketch, equal to the ionospheric transpolar current flow. The inner dotted field lines indicate the boundary of the region accessible to the data employed in this study. latitude ionospheric currents are shown by the black arrows in the figure, directed south to north at Ψ N = 90°a nd north to south at Ψ N = 270°, whose mode of closure is inaccessible to these data but likely involves further field-aligned current flow interior to the region depicted.
Similarly, for the southern PPO system in Figure 15c, sketched in Figure 16b in a similar format to Figure 16a and showing peak currents per radian on the Ψ S = 90°À 270°meridian (Figure 1h), the main field-aligned current in the southern hemisphere flowing in a~5°colatitude band from near the OCB to~20°southern colatitude (mapping to~9 R S in the equatorial plane) totals~1.3 MA rad À1 , flowing downward at Ψ S = 90°a nd upward at Ψ S = 270°. The contributions to this total current are~0.3 MA rad À1 from the southern polar ionosphere and~1.0 MA rad À1 from the closed field region at larger colatitudes. Of the total field-aligned current,~0.6 MA rad À1 (i.e., slightly under half) flows into the northern ionosphere, upward at Ψ S = 90°a nd downward at Ψ S = 270°, the difference again indicating that~0.7 MA rad À1 closes cross field in the magnetosphere. No significant additional contributions come from the northern polar ionosphere. Equatorward of these main currents, a small reversed return field-aligned current of~0.2 MA rad À1 flows in the northern ionosphere, downward at Ψ S = 90°, and upward at Ψ S = 270°, while on these field lines in the southern ionosphere a larger return field-aligned current of~0.6 MA rad À1 flows, upward at Ψ S = 90°, and downward at Ψ S = 270°. This difference indicates that of the total~0.7 MA rad À1 cross-field closure current mentioned above,~0.4 MA rad À1 closes via these return currents to the southern ionosphere (Figure 1h), again shown as two contributions of 0.2 MA rad À1 either side of the equator in Figure 16b, while the remaining 0.3 MA rad À1 (equal and opposite to the polar current) closes cross magnetosphere in the opposite magnetic meridian, again shown by the equatorial long dashed black arrows either side of the planet. An ionospheric meridional current of~0.4 MA rad À1 also remains flowing at the~21.5°southern colatitude outer limit of joint northern and southern data in both hemispheres, directed south to north at Ψ S = 90°and north to south at Ψ S = 270°, whose mode of closure is again inaccessible to these data (inner dotted field lines).
Within the limitations of accessibility, these results thus indicate that the PPO-related field-aligned currents close principally in the reversed current layers occurring separately in the Ψ N,S = 270°and 270°meridians, with relatively smaller currents,~20% of the total, being involved in the transpolar and trans-meridional flow. Of the current flowing in the meridians, roughly half closes cross field within the magnetosphere, and half via the opposite ionosphere (Figures 1e and 1h).
We note that the above discussion (and Figure 15) has been in terms of the maximal PPO-related meridional currents per radian of azimuth on the Ψ N,S = 90°À 270°meridians. However, assuming the currents vary near sinusoidally with PPO phase the values can simply be converted to total currents of a given sense by multiplying by a factor of 2. Account can also be taken of the likely reduction in quoted peak values by a factor of 2 ffiffi ffi 2 p =π À Á ≈0:90 resulting from data averaging in 90°× 90°phase squares as described in section 3.3, such that current per radian for the PPO systems can be converted into total current of a given sense by multiplying by a factor of 2= 2 ffiffi ffi 2 p =π À Á ¼ π= ffiffi ffi 2 p À Á ≈2:2.

Comparison of Results With the 2008 Data Set
We now compare our results with those derived from the 2008 data set, specifically those of Hunt et al. (2014) for the southern hemisphere data and Hunt et al. (2015) for the northern hemisphere data. We recall from Figure 3a that the 2008 data set corresponds to late Saturn southern summer with solar latitudes between À8°and À4°, while the 2012/2013 data analyzed here correspond to midnorthern spring with solar latitudes between +15°and +19°. For simplicity we will refer to the southern hemisphere in 2008 and the northern hemisphere in 2012/2013 as the "summer" hemispheres, and the northern hemisphere in 2008 and the southern hemisphere in 2012/2013 as the "winter" hemispheres.
We begin with the subcorotation currents and in Figure 17 compare the northern and southern current profiles for the 2012/2013 data set (solid lines from Figures 11a and 12a) and those derived by Hunt et al. (2014Hunt et al. ( , 2015 for the 2008 data set (crosses joined by dotted lines). Figures 17a and 17c directly compare the northern and southern hemisphere subcorotation currents, respectively, derived from the 2012/2013 and 2008 data sets.
In the 2012/2013 data there is a strong difference between the northern and southern subcorotation current profiles as discussed in section 3.6. The profiles from 2008 are seen to show a very similar asymmetry but reversed between hemispheres. The closeness of the comparison is emphasized in Figures 17b and 17d. Figure 17b again shows the northern hemisphere profile from the 2012/2013 data set but now compared with the southern hemisphere profile from 2008 mapped into the northern ionosphere (as in Figure 15). The profiles are seen to be very similar, though the 2008 data peak at slightly larger values at slightly larger colatitudes. In  Figure 17d we again show the southern hemisphere profile from 2012/2013 but now compare it with the northern hemisphere profile from 2008 mapped into the southern ionosphere. The two profiles are again seen to be very similar, including in the polar region despite the detailed seasonal difference between the two intervals, with the polar ionosphere being in permanent darkness to~6°colatitude for the 2008 data and to~14°for the 2012/2013 data. Equatorward of the OCB, the current reduces to smaller values more gradually in the winter hemisphere than in the summer hemisphere in both cases, despite the hemispheric symmetry generally expected on closed field lines. Overall, these findings clearly confirm that the hemispheric asymmetries in the subcorotation currents are related to seasonal conditions at Saturn, as suggested by Hunt et al. (2015).
We now compare the PPO-related current profiles derived from the two data sets, noting from Figure 3e that the PPO amplitudes had likely declined modestly between the two intervals. Comparisons are shown in Figure 18, where again solid lines show the profiles for the 2012/2013 data set (Figures 11 and 12), while the crosses joined by dotted lines show the profiles for the 2008 data set derived by Hunt et al. (2014Hunt et al. ( , 2015.  Hunt et al. (2014Hunt et al. ( , 2015 from the 2008 data set (crosses and dotted lines). Profiles for Ψ N,S = 90°are shown blue, while equal and opposite profiles for Ψ N,S = 270°are shown red. (a) Comparison of currents of the northern PPO system in the northern hemisphere (the corrected solid line profiles in Figure 11b) and (b) comparison of the currents of the southern PPO system in the northern hemisphere (Figure 11c), plotted versus northern ionospheric colatitude. (c) Comparison of currents of the southern PPO system in the southern hemisphere (the corrected solid line profiles in Figure 12c), plotted versus southern ionospheric colatitude. No northern PPO perturbations could be discerned in the southern hemisphere in 2008. Vertical dashed lines show the averaged OCB for each hemisphere . Figure 18a shows the northern PPO system current profiles in the northern hemisphere (the solid line corrected profile in Figure 11b for the 2012/2013 data set). The forms of the two determinations are seen to be fairly similar, though with a reversal in the sense of the current at a larger northern colatitude for the 2012/2013 data set than for the 2008 data and with a peak at a modestly smaller value at a modestly smaller colatitude. The return region of lesser current flowing at larger colatitudes beyond the peak (see Figure 1e) also appears to be sharper in the 2012/2013 data than in 2008. Figure 18b similarly compares the southern PPO system current profiles in the northern hemisphere, which are again seen to be similar in form though again differing in detail at colatitudes around the OCB and with generally smaller values for the 2012/2013 data set than for the 2008 data. Figure 18c shows the southern PPO system current profiles in the southern hemisphere (the solid line corrected profile in Figure 12c for the 2012/2013 data). The profiles from the two intervals are again similar, though again the return current region flowing at larger colatitudes beyond the peak appears to be sharper in the 2012/2013 data than in 2008. We do not show a comparison of the northern PPO system profile in the southern hemisphere in this case, since Hunt et al. (2014Hunt et al. ( , 2015 found little evidence of northern PPO modulations in the southern hemisphere in the 2008 data.

Summary
In this paper we have analyzed azimuthal magnetic field data from the third interval of highly inclined orbits of the Cassini spacecraft during 2012/2013, which allow study of the nightside perturbations associated with field-aligned magnetosphere-ionosphere coupling currents. This interval occurred during Saturn northern spring when the latitude of the Sun varied between +15°and +19°. A primary comparator data set was that obtained during the second interval of highly inclined orbits in 2008, analyzed previously by Hunt et al. (2014Hunt et al. ( , 2015, which similarly allows study of the nightside coupling currents but under late southern summer conditions when the latitude of the Sun varied from À8°to À4°. An important topic, therefore, has been a comparison of these pre equinox and postequinox data to identify seasonal or other differences. With regard to PPO conditions, previous work has shown that the southern system period was longer than the northern during both intervals but with a significantly larger separation during 2008 than during 2012/2013 (Andrews et al., 2012;Provan et al., 2013Provan et al., , 2014Provan et al., , 2016. Nevertheless, the 2012/2013 interval encompassed approximately four PPO beat periods, so that a full range of relative phases was experienced. A further difference is that the near-periapsis equatorial crossings during 2012/2013 were at larger radii (~6-8 R S ) than during 2008 (~3-6 R S ). Consequently, the 2012/2013 traversals were slower than during 2008 and do not approximate to snapshots of the current systems at fixed PPO phases, thus requiring statistical rather than a passby-pass analysis. The whole 2012/2013 data set has thus been combined into 90°× 90°"squares" of northern and southern PPO phase then used to determine colatitude profiles of the ionospheric meridional current, from which the subcorotation and PPO-related currents have been separated. Principal results are as follows.

Subcorotation Currents
In the northern (summer) hemisphere in 2012/2013 significant meridional currents are present to the smallest colatitudes observed~8°, slowly rising from~0.5 to~1 MA rad À1 at~15°(~1.5°equatorward of the OCB), then falling more rapidly to near-zero values at~17.5°. Assuming axisymmetry, the total downward currents in these regions are~3 MA in the unobserved poleward region, implying an averaged current density of 17 nA m À2 , with a further~3 MA flowing downward in the trans-OCB ring between~8°and~15°, with a lower current density of~7 nA m À2 . The total polar downward current is thus~6 MA, which returns up along the field in the auroral region ring between~15°and~17.5°, with an implied averaged current density of 27 nA m À2 . The latter value, however, is no doubt strongly reduced from typical peak values in individual cases by data averaging. The upward current region maps equatorially between~10 and~17 R S according to the field model employed here.
In the southern (winter) hemisphere the current has small (and unphysical negative) values from the poleward limit at~8°to~11°southern colatitude, rising to a 1 MA rad À1 peak just equatorward of the OCB at 16°. The total downward current in this region is thus again~6 MA, now flowing principally in a~5°ring of open field lines bordering the OCB, with an averaged current density of~16 nA m À2 . The current falls more gradually to small values on closed field lines between~17°and~22.5°, with an averaged upward current density of~10 nA m À2 . The upward current maps equatorially between~6.5 and~16 R S , thus in to somewhat smaller radial distances than the northern upward current.
Comparison of these results with those of Hunt et al. (2014Hunt et al. ( , 2015 for the 2008 data set reveals very similar polar region asymmetries but in opposite hemispheres, thus showing the effects to be seasonal. The near-zero currents at smallest colatitudes under winter conditions are taken to be due to the lack of both insolation and particle precipitation in this central open field region leading to very low ionospheric conductivities. The lack of conjugacy on closed field lines noted above, unexpected in simple theoretical models (e.g., Cowley, Bunce, & O'Rourke, 2004), was also present in opposite hemispheres in the 2008 data (Hunt et al., 2015).

PPO-Related Currents
We now summarize results concerning the northern and southern PPO current systems, discussed in terms of the maximal currents per radian of azimuth flowing on the Ψ N,S = 90°À 270°PPO meridians (as in Figure 15). Assuming that these currents vary near sinusoidally with PPO phase, however, and taking account of the likely small reduction in quoted peak values by a factor of~0.9 resulting from data averaging in 90°× 90°p hase squares (section 3.3), current per radian can simply be converted to total current of a given sense by multiplying by a factor of~2.2 (section 3.7). For the northern PPO system the main field-aligned current in the northern hemisphere flows in a~3°colatitude band whose poleward boundary lies just equatorward of the OCB and whose equatorward boundary maps to the equatorial plane at~11 R S , thus essentially collocated with the auroral upward currents of the subcorotation system. The magnitude of this maximum fieldaligned current is~1.2 MA rad À1 , flowing upward at Ψ N = 90°and downward at Ψ N = 270°, with a peak current density of~46 nA m À2 , such that the total field-aligned current of a given sense flowing in the northern hemisphere is~2.7 MA, approximately half that of the upward auroral region current of the subcorotation system, which totals to~6 MA. Where the northern PPO current flows out of the ionosphere the combined total auroral region upward current is thus approximately doubled, while where it flows into the ionosphere it is approximately canceled (Figure 11d). Similarly, for the southern PPO system the magnitude of the maximum main field-aligned current in the southern hemisphere is~1.3 MA rad À1 , flowing downward at Ψ S = 90°and upward at Ψ S = 270°, with a peak current density of~23 nA m À2 . This current is again essentially collocated with the auroral upward current of the subcorotation system and is of comparable magnitude, flowing in ã 5°colatitude band from just equatorward of the OCB to~20°southern colatitude mapping to~9 R S in the equatorial plane. For both these systems~20% of the main field-aligned current is contributed from the polar ionosphere in the corresponding hemisphere (none from the opposite hemisphere) and~80% from closed field lines at larger colatitudes, indicating that the principal component of these current systems consists of the adjacent field-aligned current layers of reversed sense flowing on closed field lines in a given PPO meridional sector (Figure 16). The cross-over point between the meridional currents of opposite senses contributed from the polar and closed field regions, indicative of a reversal in sense of the PPO-related ionospheric flow (Figure 1d), is located on closed field lines at~15°colatitude in the northern hemisphere (~16.5°colatitude in the southern hemisphere), mapping equatorially to~17 R S . Of the main PPO fieldaligned currents in both hemispheres, approximately half closes via the magnetosphere and half via the opposite hemisphere ( Figure 16). With regard to overall current closure, however, we note that~35% of the total current remains flowing in the ionosphere equatorward of the region to which these data provide access, so that the closure of this component cannot be addressed.
Comparison of the 2012/2013 PPO-related current profiles with those derived from the 2008 data shows overall similarity, though current magnitudes are generally modestly less in the 2012/2013 data set, thus reflecting similar results obtained by Provan et al. (2016) from analysis of unfiltered azimuthal field data on equatorial passes during this interval (Figure 3e). The principal difference is that the currents in both the northern and southern hemispheres appear to be similarly dual modulated in the 2012/2013 data set, with the amplitude from the opposite hemisphere being approximately half that of a given hemisphere both north and south (Figure 16), while only the northern hemisphere currents were clearly dual modulated in 2008. The return currents in the 2012/2013 data also appear to be sharper and confined to smaller colatitudes than in the 2008 data set, though the coverage of colatitude is more restricted in 2012/2013. Unlike the subcorotation currents, therefore, the differences in PPO-related currents noted here appear to be related in a more complex way to Saturn's seasons.

Appendix A: Passes in the 2012/2013 Data Set With Atypical Signatures
As indicated in section 2.1, 3 polar passes from the 2012/2013 data set, out of a total of 52, were excluded from the statistical study presented due to the presence of atypical azimuthal field and hence current signatures (as in the prior studies of the 2008 data set by Hunt et al., 2014Hunt et al., , 2015. In this appendix we exhibit these data and briefly comment on them. The data from these passes, namely, from the northern hemisphere passes on Revs 173 and 182, and the southern hemisphere pass on Rev 169, are shown in Figures A1a and A1b, respectively, in a similar format to Figures 7b and 7d, respectively. The pale blue data in each figure reproduce the I m meridional current profiles versus ionospheric colatitude θ iN,S shown in each of the latter figures (derived using the corrected azimuthal field), serving here only to indicate the range of variability in the northern and southern hemisphere data that have been included in the study. The data from the excluded passes (again using the corrected azimuthal field) are then overplotted, in red and orange in Figure A1a for Revs 173 and 182, respectively, and red in Figure A1b for Rev 169. It can be seen that the data Figure A1. Plots comparing the ionospheric meridional current profiles from passes excluded from this study (red and orange) with those which have been included (pale blue) reproduced from Figures 7b and 7d. (a) The I m profiles from the excluded northern hemisphere passes from Revs 173 (red) and 182 (orange) plotted versus northern ionospheric colatitude θ iN and (b) the I m profile from excluded southern hemisphere pass from Rev 169 (red) plotted versus southern ionospheric colatitude θ iS . The vertical black dashed line shows the averaged position of the OCB in the northern hemisphere  shown in red (northern Rev 173 and southern Rev 169) both exhibit large and atypical current signatures just equatorward of the OCB in each hemisphere, which are not matched on any of the other passes. The data shown in orange (northern Rev 182) similarly exhibit a large positive current peak at~17°whose value is around twice that of any other pass in this vicinity. These signatures are taken to be evidence of the occurrence of unusual and relatively short-lived (one~1-day pass or less) dynamic activity taking place within the magnetosphere during these passes (e.g., Bunce et al., 2010). Due to the modest size of our overall data set, inclusion of such "outlier" data risks overly influencing analysis outcomes, particularly those obtained by division of the data into phase squares as shown in Figures 9 and 10.