University College London, Holmbury St. Mary, Dorking, Surrey, RH5 6NT, UK.

Transcription

1 Cassini observations of a vortex structure in Saturn s dayside magnetosphere driven by the Kelvin-Helmholtz instability A. Masters, 1 N. Achilleos, N. Sergis, M. K. Dougherty, 1 M. G. Kivelson, C. S. Arridge,, S. M. Krimigis, H. J. McAndrews, M. F. Thomsen, S. J. Kanani,, N. Krupp, A. J. Coates., Space and Atmospheric Physics Group, The Blackett Laboratory, Imperial College London, Prince Consort Road, London, SW AZ, UK. Atmospheric Physics Laboratory, Department of Physics and Astronomy, University College London, Gower Street, WC1E BT, UK. Office of Space Research and Technology, Academy of Athens, Soranou Efesiou, Athens, Greece. Institute of Geophysics and Planetary Physics, University of California Los Angeles, Los Angeles, California, 00, USA. Mullard Space Science Laboratory, Department of Space and Climate Physics, University College London, Holmbury St. Mary, Dorking, Surrey, RH NT, UK. Centre for Planetary Sciences, University College London, London, WC1E BT. Applied Physics Laboratory, Johns Hopkins University, Laurel, MD 0, USA. Space Science and Applications, Los Alamos National Laboratory, Los Alamos, New Mexico,, USA. Max-Planck-Institut für Sonnensystemforschung, Max-Planck-Strasse, Katlenburg-Lindau, Germany. 1

2 Abstract We present the first observations of a vortex structure in Saturn s dayside, outer magnetosphere. The identification of the structure provides conclusive evidence of the operation of the Kelvin-Helmholtz (K-H) instability at Saturn s magnetospheric boundaries. Cassini observations taken during the inbound pass of the spacecraft s Revolution B orbit in Decmber 00 are analysed. Magnetic field conditions during the magnetopause crossings that occurred on this orbital pass suggest that the boundary was highly K-H unstable. Following multiple magnetopause crossings the spacecraft encountered the low-latitude boundary layer. Magnetic field, thermal plasma, and superthermal plasma observations made by Cassini during the spacecraft transition between the boundary layer and magnetosphere proper are consistent with an encounter with a vortex structure on the edge of the boundary layer this interface is also anticipated to be K-H unstable. High-energy (>0 kev) electrons observed while the spacecraft was within the vortex suggest that the structure was associated with auroral emissions. A model of the coupling between an outer magnetospheric vortex and Saturn s ionsphere via field-aligned currents is proposed. Estimates based on Knight s theory imply that field-aligned potentials of a few kv were associated with the region of upward-directed field-aligned current in the northern ionosphere, and that the resulting precipitation of accelerated electrons produced UV auroral emissions with an intensity of a few kr. We propose that K-H vortices in Saturn s outer magnetosphere produce bright spots of UV aurora. This discovery has implications for our understanding of the interaction between the solar wind and Saturn s magnetosphere.

3 1. Introduction The growth of the Kelvin-Helmholtz (K-H) instability at the boundary of a planetary magnetosphere has long been considered to be an important aspect of the solar windmagnetosphere interaction. The instability operates at a fluid interface and can manifest itself in the form of waves on the interface itself, which can then evolve into vortices as the instability enters its nonlinear phase. In a space plasma environment conditions are favourable for the growth of the K-H instability where there is a large velocity shear between plasma flows, combined with a local magnetic field that is approximately perpendicular to the relative flow direction, reducing the stabilizing effect of the magnetic tension force [Southwood, 1]. The problem of the K-H stability of a planetary magnetopause was first discussed by Dungey [1], and has since been the subject of numerous theoretical studies [Kivelson and Pu, 1, and references therein]. Spacecraft observations of Earth s magnetopause have revealed a highly structured boundary that exhibits significant spatial and temporal variability [e.g. De Keyser et al., 00]. It has been shown that the boundary is typically in motion at a speed of the order of km s -1 and that the magnetopause current layer has a variable thickness [Berchem and Russell, 1; Haaland et al., 00]; furthermore, a region of mixed magnetosheath and magnetospheric plasma on the planetward side of the magnetopause at low-latitudes has been identified. This region, hereafter referred to as the boundary layer, is characterized by tailward flows and is a quasi-permanent magnetospheric feature [Freeman et al., 1; Eastman and Hones, 1]. The boundary layer is attached to the magnetopause current layer, and its highly variable

4 thickness is generally greater than that of the current layer itself [Eastman and Hones, 1; Sckopke et al., ]. In this complicated region of interaction between the solar wind and Earth s magnetic field it has been proposed that the Kelvin-Helmholtz instability can operate at both the magnetopause and the planetward edge of the boundary layer [e.g. Lee et al., ]. Indeed, the evaluation of the K-H instability criterion at both surfaces using in situ spacecraft observations implies that the dayside of the terrestrial magnetopause is generally stable, whereas the dayside boundary layer edge is often unstable [Ogilvie and Fitzenreiter, 1]. Many of the examples of boundary waves identified on Earth s magnetopause may have been driven by the K-H instability [e.g. Owen et al., 00]; however, the most convincing evidence for the operation of the instability concerns the identification of boundary vortices. Fairfield et al. [000] compared data taken by the Geotail spacecraft during a set of crossings of the distant flank magnetopause with simulation results presented by Otto and Fairfield [000], showing that the magnetic and plasma signatures could be interpreted as spacecraft encounters with boundary vortices propagating tailward. Further convincing evidence was presented by Hasegawa et al. [00] who demonstrated that observations made by the four Cluster spacecraft in the vicinity of the equatorial, dusk flank magnetopause are also consistent with the presence of boundary vortices. The study of the growth of the K-H instability at Earth s magnetopause and boundary layer edge remains the subject of much research interest, since it has been shown to promote the transport of solar wind mass, energy, and momentum into the magnetosphere [Pu and Kivelson, 1; Miura, 1; Fujimoto and Terasawa, 1] and may provide the anomalous viscosity required to drive convection of magnetospheric plasma [Axford and Hines, ].

5 Compared to its terrestrial counterpart, Saturn s magnetosphere is far larger and possesses significant internal plasma sources, with approximate corotation with the planet being the dominant magnetospheric plasma flow. The Cassini spacecraft s orbital tour of Saturn has allowed us to study the planet s complex magnetospheric system in more detail than ever before. Figure 1 illustrates the flow of solar wind plasma around Saturn s magnetosphere, as well as the flow of magnetospheric plasma, with the structure of the magnetospheric boundary indicated. Spacecraft observations suggest that, like the case of Earth, Saturn s magnetosphere also has a boundary layer [Lepping et al., ; McAndrews et al., 00; Masters et al., 00]. The dynamics of Saturn s magnetopause include global expansion and contraction associated with variations in the dynamic pressure of the solar wind [Arridge et al., 00] and oscillation of the boundary at approximately the planetary rotation period [Clarke et al., 00; Clarke et al., 00, in preparation]. In addition, the detailed anlaysis of Cassini data taken during magnetopause crossings has revealed evidence for dayside reconnection [McAndrews et al., 00]. Given the plasma flows shown in Figure 1 it is clear that a dawn-dusk asymmetry in the flow shear across the magnetospheric boundary region exists. As a result the dawn flank boundaries are expected to be highly susceptible to the growth of the K-H instability, whereas the dusk flank boundaries are expected to be stable on the dayside [Pu and Kivelson, 1; Galopeau et al., 1]. Waves on Saturn s dawn flank magnetopause were first identified by Lepping et al. [] using Voyager 1 observations; these authors suggested that the waves may have resulted from the growth of the K-H instability. Masters et al. [00] carried out an analysis of data taken by Cassini during crossings of the magnetopause in the pre-dawn sector, which

6 also revealed evidence for boundary wave phenomena. They concluded that the most plausible driving mechanism responsible for the wave activity was the K-H instability. The growth of the K-H instability at Saturn s magnetospheric boundaries is also apparent in the results of three-dimensional (-D) magnetohydrodynamic (MHD) simulations of Saturn s magnetosphere: Fukazawa et al. [00a] reported that largescale vortices form in their simulations for a range of interplanetary magnetic field (IMF) conditions. Fukazawa et al. [00b] showed that for the case of northward IMF the boundary vortices formed on the dawn side, followed by the dusk side, in the simulations. They concluded that the vortices were similar to structures expected to result from the operation of the K-H instability, and suggested that they induced local magnetic reconnection. It has also been suggested that K-H perturbations of Saturn s outer magnetospheric environment can produce auroral emissions. Galopeau et al. [1] proposed that the growth of the K-H instability at the magnetopause can explain why the most intense source regions of Saturn kilometric radiation (SKR) Saturn s auroral radio emissions are on the morning side at high-latitudes, since the highly K- H unstable morning magnetopause should produce MHD waves capable of accelerating electrons towards the planet. Furthermore, Fukazawa et al. [00a] found that the strongest field-aligned currents (FACs) generated in their simulations were the result of vortex formation, suggesting that such structures strongly perturb magnetic field lines. These simulation results suggest that vortex structures can produce an auroral signature. A relationship between bright spots of ultraviolet (UV) auroral emission at Earth and K-H vortices in the outer terrestrial magnetosphere has previously been proposed [Lui et al., 1; Vo and Murphree, 1].

7 In this paper we address the following open issues by analysing data taken by the Cassini spacecraft during Saturn orbit: 1. Does the growth of the Kelvin- Helmholtz instability at Saturn s dawn flank magnetospheric boundaries lead to vortices on the dayside?. If vortices form, are they capable of generating current systems that produce auroral emissions? In Section we examine data taken during the boundary crossings made on the inbound pass of a Cassini orbit that took place in December 00. The choice of orbit was based on our expectation of the conditions under which the morning magnetopause boundary was most K-H unstable. We show that an atypical magnetic and plasma transition from the boundary layer to the magnetosphere proper took place during the inbound pass of the orbit. In Section we analyse the data taken during the transition, and interpret it as a spacecraft encounter with a K-H vortex structure on the edge of the boundary layer. We then demonstrate that this interpretation in consistent with the observations. In Section we discuss the coupling between the vortex and the ionosphere, addressing the question of whether or not an auroral signature was produced. Finally, in Section we summarise our results and re-define the open issues concerning this topic Observations The coordinate system utilized throughout this study is the Kronocentric Solar Magnetospheric (KSM) system, which is Saturn-centered with the positive x axis pointing toward the Sun. The z-axis is chosen such that the x-z plane contains Saturn s magnetic dipole axis, with the positive z-axis pointing north. The y-axis completes the orthogonal set, with the positive y-axis pointing towards dusk. The unit of distance used is Saturn radii (R S ; 1 R S = 0, km).

8 The Cassini spacecraft has been in Saturn orbit since 1 July 00. The instruments mounted on the three-axis stabilized orbiter allow us to study the planetary magnetospheric environment in detail [Blanc et al., 00]; in this paper we present data taken by three of these orbiter-mounted instruments. The first instrument of interest is the dual technique magnetometer (MAG) [Dougherty et al., 00], which provides measurements of the local magnetic field vector. MAG data used in this study was taken by the fluxgate magnetometer sensor at 1 s resolution unless otherwise stated. The second instrument of interest is the Cassini plasma spectrometer (CAPS) [Young et al., 00], which allows us to examine the thermal plasma properties. Data taken by two of the CAPS sensors are used: the electron spectrometer (ELS), which detects electrons with energies between 0. ev and. kev; and the ion mass spectrometer (IMS), which detects ions with energies between 1 ev and 0. kev per charge. The time resolution of ELS (IMS) data presented here is 1 min ( s). The third instrument of interest is the magnetospheric imaging instrument (MIMI) [Krimigis et al., 00], which measures high-energy particles. Data taken by the low energy magnetospheric measurements system (LEMMS) of the MIMI instrument are presented, which detects electrons with energies between ~0 kev and ~ MeV. -min time resolution LEMMS data is used in this study. During the initial phase of Cassini s orbital tour the spacecraft explored the low-latitude dawn side magnetosphere, regularly crossing both the bow shock and magnetopause boundaries [e.g. Dougherty et al., 00]. We evaluated the K-H stability of the morning magnetopause during each orbital pass by comparing the typical magnetic fields measured immediately before and after each magnetopause crossing, since the magnetic field vector was consistently measured on both sides of the boundary and affects the growth of the instability. Compared to the fluctuating

9 magnetosheath magnetic field, the magnetospheric field was typically steady and predominantly southward immediately inside the magnetopause on each pass as expected. This southward orientation is approximately perpendicular to the anticipated magnetosheath and magnetospheric flows in the equatorial region, which are approximately anti-parallel. As mentioned in Section 1, magnetic tension forces act to stabilize the boundary; thus the ideal conditions for growth of the K-H instability at Saturn s magnetopause correspond to parallel or anti-parallel magnetic fields at the interface that are perpendicular to the relative flow, where the wave vector of the fastest growing K-H wave mode is aligned with the relative flow vector [Kivelson and Pu, 1]. Under such conditions the growth rate of the K-H instability increases, making the formation of vortex structures on the dayside more likely. Consequently, we selected the orbital pass where the average magnetosheath and magnetospheric magnetic fields were closest to being parallel or anti-parallel for further examination: the inbound pass of the Revolution (Rev) B orbit in December 00. Figure shows the spacecraft trajectory during part of the inbound pass of the Rev B orbit. The first magnetopause crossing took place at a range of. R S, a magnetic latitude of -1.º, and a Saturn local time (SLT) of 0. Figure shows Cassini observations made on 1 December 00 during a h period encompassing the boundary crossings. Data taken by the MAG, ELS, and LEMMS instruments are shown, revealing the properties of the local magnetic field and electrons over the combined ELS-LEMMS energy range. During the period shown in Figure the CAPS sensors were actuating. This periodic sweeping of the sensors led to the periodic modulation seen in the ELS time-energy spectrogram. Throughout the period a spacecraft photoelectron population was observed by ELS at energies below ~ ev.

10 Due to pointing constraints the derivation of ion moments from IMS data was not possible during the interval; however, an IMS time-energy spectrogram is shown later in this section to provide evidence for variations in the plasma bulk flow direction. The spacecraft began the day in the magnetosheath before making three magnetopause crossings and ending the period shown in Figure in the magnetosphere. Such multiple crossings during an orbital pass are the result of boundary motion at speeds greater than that of the spacecraft. Figure c shows the magnetic field elevation angle, which is defined as the angle between the field vector and the positive z direction in KSM coordinates, with 0º corresponding to a field vector in the positive z direction, 0º corresponding to a field vector lying in the x-y plane, and -0º corresponding to a field vector in the negative z direction. This elevation angle clearly shows that throughout the different magnetic regimes the magnetic field stayed predominantly southward, revealing a low magnetic shear across the magnetopause the angular difference between the typical magnetosheath and magnetospheric fields was <º. The first magnetopause crossing occurred at a universal time coordinate (UTC) of ~0:0. During the crossing there was an increase in the magnetic field strength, a decrease in the electron number density, and an increase in the electron temperature. This behavior is typical of inbound crossings of Saturn s magnetopause [McAndrews et al., 00; Masters et al., 00]. In the following magnetosphere excursion (~0:0 to ~0:0 UTC) a significantly higher intensity of energetic (>0 kev) electrons was observed by LEMMS compared to the preceding magnetosheath interval, with pitch angle information revealing that the high-energy electron distribution was principally field-aligned and bi-directional. These electron distributions have been previously reported by Saur et al. [00], who concluded that they were the result of auroral

11 processes. The ELS electron distributions measured during this interval do not display any clear pitch angle anisotropy, although the existence of such an anisotropy cannot be ruled out since it may not have been able to be resolved given the sensitivity level of the ELS sensor. During the magnetosphere excursion two electron number density increases occurred, which were correlated with electron temperature decreases. Given that the bi-directional, superthermal electrons were also detected during these features, we identify them as excursions into the boundary layer. Such transient encounters with the boundary layer can be explained by spatial or temporal variations of the boundary layer itself, motion of the boundaries relative to the spacecraft, or a combination of these effects. The second magnetopause crossing occurred at ~0:0 UTC. This outbound crossing into the magnetosheath was associated with an electron number density increase and temperature decrease as expected; however, the magnetic field magnitude did not decrease to values similar to that measured in the magnetosheath prior to the first crossing. Instead, during the subsequent magnetosheath excursion (~0:0 to ~0:00 UTC) the field magnitude remained comparable to that observed during the preceding magnetosphere excursion, including a number of intervals with a duration of the order of minutes where the field magnitude was greatly reduced. We note that the electron number density measured during this magnetosheath excursion was generally ~0% lower than that measured in the earlier magnetosheath interval. The distinct high-energy electron distributions were no longer observed, indicating that the spacecraft was no longer in a region of closed field lines. We identify the region of the magnetosheath sampled by the spacecraft during this magnetosheath excursion as a plasma depletion layer immediately upstream of the magnetopause [e.g. Zwan and Wolf, 1]. In such regions the draping of IMF field

12 lines around the magnetospheric obstacle results in an increased field magnitude associated with a decrease in the plasma density and temperature [Phan et al., 1]. Zwan and Wolf [1] suggested that such depletion layers form when dayside reconnection is unimportant, which is in agreement with the almost parallel magnetosheath and magnetospheric magnetic fields observed by Cassini on 1 December 00, since such magnetic conditions are not favorable for low-latitude dayside reconnection. We note that the intervals of depressed magnetic field magnitude that occurred during the excursion were coincident with intervals of elevated electron number density, implying that these features could correspond to mirror mode waves in the depletion layer produced by the operation of the mirror instability, which grows under conditions of high plasma beta and ion temperature anisotropy [e.g. Tsurutani et al., 1]. Assuming that the solar wind plasma was composed of protons and electrons, and that the proton and electron number densities and temperatures were equal, the plasma beta during the excursion was ~. Since we are unable to measure the ion temperature anisotropy we cannot confirm that this magnetosheath region was unstable to the growth of the mirror instability when it was traversed by Cassini. Violante et al. [1] showed that the observations made by the Voyager spacecraft during their Saturn fly-bys also suggest the presence of a plasma depletion layer and mirror mode waves in the planetary magnetosheath. At ~0:00 UTC Cassini made its third and final magnetopause crossing; the plasma signature of this crossing was less distinct than that of the two preceding crossings. Between ~0:00 and ~0:0 UTC the electron number density gradually decreased and the electron temperature gradually increased. During the same interval an increase in the intensity of high-energy electrons was observed by LEMMS, and, 1

13 1 1 as for the high-energy electrons detected in the earlier magnetosphere excursion, the particle pitch angles reveal that these high-energy electrons were principally fieldaligned and bi-directional. This suggests that the spacecraft was located on closed field lines from ~0:00 UTC onwards. We surmise that the spacecraft crossed the magnetopause at that time and was subsequently immersed in the boundary layer, where the plasma properties gradually became less similar to those of the magnetosheath as the distance between the spacecraft and the magnetopause increased and the spacecraft approached the edge of the boundary layer [Eastman and Hones, 1]. Following this boundary layer excursion Cassini made a disturbed transition into the magnetosphere proper. The vertical dashed lines in Figure indicate the duration of this unusual transition. In the following section we examine the data taken by Cassini during this transition Vortex encounter Data taken by MAG, ELS, IMS, and LEMMS during a h interval encompassing the boundary layer-magnetosphere proper transition are shown in Figure. The interval is divided into sub-intervals 1 through. During sub-interval 1 the spacecraft was still in the boundary layer. During sub-interval Cassini was immersed in a hotter, more tenuous electron environment, coincident with a strikingly higher intensity of superthermal electrons detected by LEMMS. The vast majority of the pitch angles of these high-energy particles were between 1 and, thus these electrons were predominantly traveling anti-parallel to the local magnetic field (in the northward direction). During sub-interval Cassini was immersed in a colder, denser electron 1

14 environment, similar to that encountered during sub-interval 1; however the distinctive high-energy electron distribution persisted. Finally, during sub-interval the spacecraft was in the magnetosphere proper, where the lowest electron number densities and highest electron temperatures encountered during the overall interval shown in Figure were measured. The intensity of superthermal electrons was significantly lower than that of the previous two sub-intervals, and the uni-directional distribution was no longer present. We propose that this unusual transition between Saturn s boundary layer and magnetosphere proper was caused by the spacecraft encountering a vortex structure on the edge of the boundary layer driven by the growth of the K-H instability. Given that the flow directions in the boundary layer [Freeman et al., 1] and magnetosphere proper [Richardson, 1] are expected to be approximately antiparallel, and that there were almost parallel, southward magnetic fields either side of the interface, the growth of the K-H instability at the boundary layer edge at this time is anticipated [e.g. Kivelson and Pu, 1]. The structure of the vortex encountered by Cassini and the passage of the spacecraft through it that we propose are illustrated in Figure. In our interpretation the spacecraft was within the vortex during sub-intervals and, with the environment sampled by Cassini during these sub-intervals corresponding to a region of deflected magnetospheric plasma and a region of deflected boundary layer plasma, respectively. Clearly the alternating regions of relatively cold and dense and relatively hot and tenuous plasma can be explained by a spacecraft encounter with a K-H vortex. We would expect the vortex structure to propagate tailward at a speed greater than that of the spacecraft, and lead to a twisted magnetic field topology [Fukazawa et 1

15 al., 00a, b]. To examine the field perturbations we subtracted the average field for the entire interval shown in Figure from min resolution field measurements, producing a set of field perturbation vectors. In Figure d the projections of these perturbation vectors into the M-N plane of the boundary normal coordinate system are shown. The boundary normal coordinate system is a right-handed, orthogonal system with axes labelled L, M, and N [Russell and Elphic, 1]. The N axis points along the boundary normal direction (directed away from the planet) and the L axis is chosen such that the magnetospheric magnetic field lies in the L-N plane, with the M axis completing the set. For the case of the 1 December observations of the boundary layer edge we have used an N direction given by the prediction of the Arridge et al. [00] model of Saturn s magnetopause surface, assuming that the unperturbed orientation of the boundary layer edge is the same as that of the magnetopause. The average field for the entire interval shown in Figure was taken as the nominal magnetospheric field vector, used to defne the L axis. The resulting M direction points roughly towards the subsolar point. Thus in Figure d the field perturbation vectors have been projected into a plane that is approximately perpendicular to the typical orientation of the magnetospheric magnetic field. The N direction has been defined as pointing down, and the M direction has been defined as pointing to the right. Therefore we are viewing the projections from the North, looking along the nominal field orientation, with down pointing into the magnetosheath and right pointing towards dusk along the suface. The schematic shown in Figure illustrates the proposed cross-sections of the boundaries in the same M-N plane. 1

16 There was a clear, systematic variation of the perturbation vectors during the transition. The largest field perturbations (~ nt) were measured during sub-intervals 1 and, before and after the spacecraft was inside the vortex structure respectively. During sub-interval the field perturbation vectors all had a negative N component, whereas during sub-interval they all had a positive N component. During both subintervals and the maximum field perturbation was ~1 nt. This signature suggests that a boundary perturbation propagated past the spacecraft and resulted in the variations in the thermal plasma conditions. Furthermore, the positive-negative variation of the N component reveals that the disturbance was moving approximately parallel to the M direction, along the boundary layer edge, also in agreement with the propagating boundary perturbation hypothesis. The approach we have used to examine the magnetic field perturbations during the encounter was also employed by Hasegawa et al. [00] in their analysis of magnetic field measurements made by each of the Cluster spacecraft during encounters with vortices on Earth s magnetopause. The behavior of the magnetic field observed by Cassini on 1 December 00 is remarkably similar to that reported by these authors; however, we note that such a magnetic field signature does not allow us to unambiguously conclude that the disturbance was a boundary vortex, since a similar signature could be produced by a boundary wave. The additional evidence required to confirm that Cassini encountered a vortex concerns the nature of the plasma flow. A time-energy spectrogram of ion energy from IMS detector is shown in Figure g. During the period of the transition the actuation of the IMS sensor led to detector periodically pointing into the expected direction of corotation: approximately the negative M direction. As mentioned in section, pointing 1

17 constraints prohibited the derivation of reliable ion moments from IMS data. Nonetheless, time-energy spectra can still be used to draw qualitative conclusions regarding the flow direction. During sub-interval 1 part of the ion distribution was observed by IMS detector ; however the same was true of all the IMS detectors. The highest countrates were measured by detector, which was closer to pointing into the anticipated direction of tailward flow than any other detector. This is consistent with tailward flow in the boundary layer region sampled during sub-interval 1. Detector did not observe clear evidence of flow in the corotation direction during sub-interval, whereas during sub-interval the flow was once again partially observed. In fact, during sub-interval higher count-rates were observed by detector than any other IMS detector, which suggests that the flow was approximately in the corotation direction. During sub-interval the flow continued to be observed by detector with higher count-rates than other IMS detectors, also suggesting approximate corotational flow. Throughout sub-intervals and the measured spectra reveal a double-peaked distribution. Our vortex encounter proposal predicts that the plasma flow was approximately parallel to the unperturbed boundary layer edge in the tailward direction during sub-intervals 1 (boundary layer) and (deflected magnetosphere proper), whereas it was approximately in the corotation direction during sub-intervals (deflected boundary layer) and (magnetosphere proper), as illustrated in Figure. Therefore the IMS data are consistent with our proposal. This information allows us to differentiate between the cases of a boundary wave and a boundary vortex, since a tailward-propagating wave interpretation cannot explain the flow approximately in the corotation direction observed during the boundary layer excursion proposed to result 1

18 from the wave motion (sub-interval ). An encounter with a vortex on the edge of the boundary layer is able to explain the flow characteristics we have deduced. The double-peaked ion distributions observed by IMS during sub-intervals and may have been due to the presence of water group ions in addition to lighter ion species. Although the time-of-flight capability of IMS does not confirm that such heavier ions were present the data does not rule out this possibility. McAndrews et al. [00] showed that such ion populations are present in Saturn s magnetosphere at ranges similar to that of the spacecraft at the time of the vortex encounter presented here. In addition, observations [Hasegawa et al., 00] and simulations [Fujimoto and Terasawa, 1] of K-H vortices at Earth s magnetopause have shown that such structures result in plasma mixing, providing a mechanism for the transport of solar wind plasma into the terrestrial magnetosphere. The particle distributions measured by Cassini during the vortex encounter may thus correspond to a complex mixed boundary layer-magnetosphere proper populations. During this period an enhancement in the electron number density was observed by ELS at ~0: UTC, which was coincident with an increase in the IMS detector count-rates and a double-peaked ion distribution. This dense feature does not appear to be directly associated with the vortex encounter as the magnetic field data does not suggest that the feature was the result of a propagating disturbance on the boundary layer edge, and the distinctive high-energy electron were not detected by LEMMS. The evidence of flow in the corotation direction implies that it was not a boundary layer excursion caused by uniform motion of the boundary region relative to the spacecraft, as tailward flow would be expected for such a case. Similar dense plasma features in Saturn s outer magnetosphere were observed by the Voyager spacecraft and discussed by Goertz [1], who suggested that they were flux tubes 1

19 that had detached from the plasma sheet due to a centrifugally driven flute instability. It is also possible that such features are related to K-H boundary vortices. It has been shown that reconnection occurs in K-H vortices on Earth s magnetopause [Nykyri et al., 00] and simulations suggest that this process results in the injection of regions of magnetosheath plasma into the magnetosphere [Nakamura and Fujimoto, 00]. The detailed examination of outer magnetospheric plasma distributions measured by Cassini should be the subject of a future study. Since there is no evidence of a magnetosheath excursion during the transition, and the presence of bi-directional, high-energy electrons indicates that the spacecraft was located on closed magnetic field lines, it is unclear whether the magnetopause boundary was perturbed at the time of the encounter. There are four possible scenarios: 1. A vortex formed on the boundary layer edge and the magnetopause was relatively unperturbed;. A vortex formed on the boundary layer edge and waves were present on the magnetopause;. Vortices formed independently on both the magnetopause and the boundary layer edge;. The vortex comprised both boundaries. Given that the Rev B orbit inbound pass was chosen because of the steady, southward IMF orientation that should produce a highly K-H unstable magnetopause, we expect that the magnetopause surface was perturbed at the time of the encounter. In Figure the magnetopause boundary is shown to exhibit a wave-like perturbation; this is purely speculation and corresponds to the second scenario. We conclude that the unusual transition between the boundary layer and the magnetosphere proper observed by Cassini on 1 December 00 was caused by the spacecraft encountering a vortex structure on the boundary layer edge resulting from the growth of the K-H instability. This interpretation is supported by the fact that the nominal conditions observed either side of the boundary layer edge suggest that the 1

20 1 1 interface was highly K-H unstable, and is in agreement with the in situ measurements made by Cassini during the transition. However, there is another feature of the vortex encounter that must be explained: the high-energy electrons detected by LEMMS. As we have seen, while Cassini was within the structure LEMMS observed a dramatic increase in the intensity of superthermal electrons some with energies as high as 00 kev with the vast majority of these particles travelling northward within 1 of the direction antiparallel to the magnetic field. Similar high-energy electron observations have been reported by Saur et al. [00] and Mitchell et al. [00], who suggested that they were associated with auroral processes. The detection of these energetic electrons while the spacecraft was within the vortex strongly suggests that the structure was associated with an auroral signature. In the following section we investigate the auroral implications of the vortex in more detail Vortex-ionosphere coupling In order to understand the coupling between the vortex structure in Saturn s outer magnetosphere and the planetary ionosphere we must have a clear picture of the vortex-induced current system. In the following analysis we use a simple approach to examine the nature of the system. Consider the initial condition of unperturbed magnetic field lines in the equatorial region of the outer magnetosphere where a K-H vortex forms. Prior to the generation of the vortex the field lines all point southward and are uniformally distributed; however, as the vortex evolves and plasma begins to circulate the magnetic field topology will also be affected due to the frozen-in flux approximation. 0

21 This results in a twisting of the magnetic field lines which will produce a current system, as predicted by simulations [Fukazawa et al., 00a]. This simple picture is shown in Figure. Field lines in the foreground and the background are differentiated to illustrate the sense of plasma circulation within the vortex, and the resulting field line perturbations. The location of the vortex structure is also indicated we note that K-H boundary vortices have complex -D structure [Takagi et al., 00], in this simple picture the vortex region represents the site where the main twist is applied to the field lines due to plasma circulation, which will affect the field lines at higher latitudes. By Ampère s law, the twisted field topology leads to electric currents that flow toward the vortex region from the North and from the South. Furthermore, in the vicinity of the vortex the nature of the field line perturbations results in current directed away from the centre of the vortex, to the edges of the structure. Current conservation can be achieved via FACs at the edges of the vortex region. These currents are indicated in Figure. The j cross B forces will act to resist the twisting of the magnetic field produced by the formation of a vortex; thus there will be a competition between these forces and the K-H instability, which derives its free energy from the kinetic energy of the plasma flows. Using the MAG observations made during the 1 December 00 vortex encounter we can assess whether the current system predicted by this simple picture was present. By examining the magnetic field perturbation vector projections shown in Figure d and applying Ampère s law we can examine the nature of the currents associated with the vortex. Since the perturbation vectors have been projected into the field-perpendicular plane (approximately the equatorial plane) and the projections are viewed from the positive L axis (looking down field lines from the North) the behavior of these vectors reveals southward FAC inside the structure, flanked by 1

22 northward FAC at the peripheries. This is consistent with the prediction of our simple picture for the case of a spacecraft passing through the vortex structure northward of the site where the main twist was applied to the field lines (see Figure ). Given this observational support for the nature of the local current system generated by the formation of a K-H vortex on the edge of the boundary layer, we propose the global current system shown in Figure, where the FACs close in Saturn s northern and southern ionosphere. This simple model predicts that the vortex footprint in the ionosphere in each hemisphere should be coincident with a region of upward-directed current, surrounded by a region of downward-directed current. These currents would likely be carried by electrons moving into and out of the ionosphere respectively. Unfortunately, imaging of Saturn s aurorae was not carried out at the time of the 1 December vortex encounter, prohibiting us from conclusively answering the question of whether or not the structure produced an auroral signature. Nonetheless, we can perform estimations based on Cassini observations to draw tentative conclusions. The spacecraft travelled 0. R S while inside the vortex (beginning of subinterval to end of sub-interval in Figure ). In the absence of bulk flow measurements we cannot determine the velocity of the structure, thus we assumed that the distance travelled by Cassini during the encounter was approximately equal to the scale size of the vortex. Using the MAG data taken during the encounter and this approximate scale size we applied Ampère s law to estimate the current density of the southward FAC within the central region of the structure to be ~0.0 na m -. As shown in Figure, this region of southward FAC observed by Cassini northward of the site of the main twist maps to the region of upward-directed FAC in the northern ionosphere.

23 Assuming that the FAC associated with a magnetic flux tube was conserved, a model of the magnetospheric field can be used to infer the current density in the vicinity of the ionosphere. Since an approximately dipolar field was measured in the magnetosphere immediately after the vortex encounter we used a dipole model of the planetary magnetic field [Dougherty et al., 00] to calculate the magnetic field strength at the altitude of Saturn s auroral zone [Gérard et al., 00], leading to an auroral zone FAC density of ~ na m -. Using this model the field lines that thread the vortex structure map to a latitude of ~, and the footprint has a latitudinal extent of less than 1. This latitude is within the range of values at which Saturn s UV auroral emissions have been observed [Clarke et al., 00]. We note that the estimated auroral zone current density is not sensitive to the choice of magnetospheric field model used; thus, since we are not examining the exact location of the footprint in detail (i.e. we are not comparing the prediction with imaging data) and instead are interested in the implications for precipitating particle energy fluxes, the use of a dipolar field model is appropriate for our purposes. The auroral zone current density at the vortex footprint that we have estimated is of the same order of magnitude as the values deduced by Cowley et al. [00] and Bunce et al. [00] for the main auroral oval. In their study of the origin of Saturn s UV aurora, Bunce et al. [00] used Cassini magnetic field and plasma observations taken while the spacecraft was at high-latitudes to identify the FAC systems associated with the southern auroral oval. They inferred that the FAC density associated with the auroral oval was ~ na m -, and employed Knight s theory [Knight, 1] to calculate the magnitude of the resulting field-aligned potential and energy flux deposited into the auroral zone. They concluded that the FAC structures produced emission with a intensity of a few tens of kr, consistent with the observed

24 intensity of the southern auroral oval. Although it has been suggested that Knight s theory is not appropriate for use in a magnetosphere with centrifugally confined plasma like Saturn s [Ray et al., 00], we assumed that the Knight relations could be used to give an indication of whether electron accelaration into the auroral zone occurred at the ionospheric footprint of the vortex encountered by Cassini on 1 December 00. In the auroral zone the estimated current density of ~ na m - that resulted from the formation of the vortex was carried by magnetospheric electrons. If the current density was greater than a critical value that could be supported by the electron population then field-aligned electric potentials would have resulted that accelerate electrons, increasing the current density to the required value. This acceleration mechanism would have led to high-energy electrons travelling away from the auroral zone (downward FAC) and into the auroral zone (upward FAC), producing auroral emissions in the latter case. The minimum field-aligned potential (Φ ) predicted by Knight s kinetic theory is given as 1 1 eφ = W e j I j e 1 (1) 1 1 and the precipitating accelerated electron energy flux (E I ) is given as 0 1 E I = E e j I j e +1, ()

25 where e is the elementary charge, W e is the thermal energy of the unaccelerated source electron population, j I is the required ionospheric FAC density, j e is the FAC density provided by the unaccelerated source electrons, and E e is the energy flux of the unaccelerated source electron population [Knight, 1; Bunce et al., 00]. This approach is based on the assumption that the source population is isotropic. Bunce et al. [00] found that the parameters of the source electrons were highly variable. Assuming typical values based on the results presented by these authors (We ~1 kev, j e ~0 A m -, E e ~0.0 mw m - ) and applying Knight s theory for the case of the upward-directed FAC generated in the northern ionosphere by the formation of the 1 December K-H vortex, we estimate that a field-aligned potential of ~ kv was produced, with a precipitating electron energy flux of ~0. mw m -. If we equate 0.1 mw m - to 1 kr of UV emission [e.g. Bunce et al., 00] then these calculations suggest that the vortex produced a UV auroral signature with an intensity of ~ kr in the northern hemisphere. Clearly these values resulting from the application of Knight s theory should be treated as order of magnitude estimates. In our calculation of the current density local to the vortex we have not been able to address spatial and temporal variations of the vortex structure or the speed of the structure in our calculation the observations suggest the vortex speed was greater than that of the spacecraft. Furthermore, we note that Bunce et al. [00] showed that the properties of the electron populations on high-latitude field lines at ~1 R S from the planet were subject to extreme variation. Using the full range of parameter values, rather than chosen typical values, the outcome of applying the Knight relations ranges from suggesting that the vortex produced no auroral signature to sugesting that the vortex produced an auroral signature far more intense than the main auroral oval, which has a typical intensity of

26 tens of kr. We must also consider that any vortex-induced signature would lie close to the main oval, assuming that the location of the main oval corresponds to the boundary between open and closed field lines [Cowley et al., 00], and that the size of the signature could have a smaller latitudinal extent. The vortex signature may also have been less intense than the main auroral emission. However, given that simulations suggest that such vortices generate strong FACs that produce discrete aurorae [Fukazawa et al., 00a], we propose that the K- H vortex encountered by Cassini on 1 December 00 produced a UV auroral signature. We note that the nature of the structure and the local time at which it was observed suggest that it was at a relatively early stage of its evolution; after the Cassini encounter we would expect the vortex to have accelerated tailward along the dawn flank, with its scale and the vortex-induced perturbation of the magnetic field increasing with time. As mentioned in Section 1, further support for our proposal is the terrestrial precedent: Lui et al. [1] and Vo and Murphree [1] presented observations of bright spots of auroral emission at Earth, which they attributed to the growth of the K-H instability at the outer magnetospheric boundaries. The motivation behind our examination of the coupling between the vortex and Saturn s ionosphere was the detection of high-energy electrons by Cassini while the spacecraft was inside the structure. According to the coupling system we have developed in this section these electrons were accelerated away from the ionosphere by field-aligned potentials, and were subsequently detected by the spacecraft in the equatorial region at ~1 R S. We note that the fact that the maxima in the intensity of high-energy electrons detected by LEMMS occurred towards the peripheries of the vortex encounter; this is consistent with our auroral acceleration mechanism and associated vortex-induced current system, as these electrons would have been

27 accelerated along field lines that map to the edges of the structure in our model. Although observations of vortices on Earth s magnetopause have been shown to lead to magnetic reconnection due to the strongly perturbed magnetic field topology [Nikutowski et al., 00; Nykyri et al., 00], the observed magnetic field during the Cassini encounter with a vortex at Saturn does not suggest the operation of the reconnection process; thus it does not appear that vortex-induced reconnection can explain the energetic electrons. The clear pitch angle asymmetry in the LEMMS electron distributions dicussed in Section must also be reconciled with this auroral processes interpretation. As the electrons were predominantly travelling northward during the Cassini encounter this suggests that they originated from the southern auroral zone and were accelerated away from the ionosphere by field-aligned potentials coincident with the regions of downward FAC. During Cassini s nominal mission (1 July 00 1 July 00) southern Summer conditions prevailed, as Saturn s near-aligned rotation and magnetic dipole axes [Dougherty et al., 00] were tilted anti-sunward. At the time of the vortex encounter the title angle was.. Auroral imaging that has been carried out during these southern Summer conditions suggests that the southern UV aurora is brighter than the northern equivalent [Esposito et al., 00], implying greater particle precipitation in the South that may be the result of larger field-aligned potentials related to higher ionospheric conductivity due to the difference in solar illumination. Talboys et al. [00] recently showed that the main auroral current systems are generally stronger in the southern hemisphere than in the northern hemisphere for southern Summer conditions. This apparently seasonal feature of Saturn s aurorae provides an explanation for the pitch angle asymetry of the high-energy electron distributions observed by

28 Cassini during the 1 December vortex encounter. Greater field-aligned potentials, associated with larger currents into and out of the southern auroral zone, could have led to the anti-planetward acceleration of electrons to higher energies in the South than in the North. The lower energy electrons accelerated away from the northern auroral zone (travelling southward in the equatorial outer magnetosphere where the spacecraft was located) may not have been clearly observed by LEMMS due to the lower limit of the LEMMS energy range, and possibly was not resolved by ELS as the signal was below the sensitivity of the sensor.. Summary In this paper we have presented and analysed observations taken by the Cassini spacecraft during the inbound pass of its Rev B orbit in December 00. This orbital pass was selected because magnetic field observations suggested that Saturn s morning magnetopause was likely to be highly unstable to the growth of the K-H instability at the time of the spacecraft crossings of the boundary. Multiple magnetopause crossings occurred during the pass on 1 December 00, with a clear boundary layer region observed following the final crossing. We have demonstrated that the spacecraft subsequently encountered a vortex on the planetward edge of the boundary layer, which was associated with a twisted magnetic field topology, highenergy (>0 kev) electrons travelling predominantly anti-parallel to the field, and deflection of the bulk plasma flow. Using a simple model of the vortex-induced magnetic field perturbations, which is consistent with the magnetic field observations, we have discussed the global current system resulting from the formation of the vortex. This current system closes

29 in the ionosphere, producing regions of upward and downward-directed FACs. Based on the results of the application of Knight s theory to investigate the vortexionosphere coupling, we propose that the vortex was associated with a UV auroral signature with an intensity of a few kr. This discovery of a K-H vortex in Saturn s dayside magnetosphere has implications for our understanding of the physics of the outer magnetosphere, and solar wind influences on Saturn s space environment and auroral emissions. Our analysis confirms the predictions of theory [Pu and Kivelson, 1] and simulations [Fukazawa et al., 00a, b] that the dawn flank magnetospheric boundaries are highly K-H unstable, and that vortex structures likely generate strong FACs that lead to auroral emissions; however, there are a number of open issues concerning this topic. In particular the following questions remain to be answered: 1. Do Cassini observations confirm the anticipated dawn-dusk asymmetry in boundary perturations caused by the growth of the K-H instability?;. What is the extent of the transport of solar wind mass, energy, and momentum into Saturn s magnetopshere?;. Does magnetic reconnection occur inside K-H vortex structures?;. Are the magnetopause and boundary layer edge independently unstable boundaries?;. Can bright spots of UV auroral emission be attributed to the presence of outer magnetospheric vortices?. What is the relationship between the K-H phenomenon and the properties of SKR emissions? To address these issues a detailed study of Cassini magnetic field and plasma data taken during magnetospheric boundary crossings combined with an examination of auroral imaging results is required. In addition, further work must be carried out in order to better understand Saturn s boundary layer and outer magnetosphere; a comparison with the equivalent regions of the terrestrial and jovian magnetosperes

30 may benefit such investigations. The observations presented in this paper hint at the potential importance of the growth of the K-H instability at Saturn s outer magnetospheric boundaries for the dynamics of the magnetosphere, making this topic worthy of further research. 0

31 Acknowledgements AM acknowledges useful discussions with D. G. Mitchell, R. J. Walker, K. Nykyri, and E. M. Henley. We acknowledge N. Powell for artwork preparation. We acknowledge the support of the MAG and MIMI data processing/distribution staff, and L. K. Gilbert and G. R. Lewis for ELS data processing. This work was supported by UK STFC through the award of a studentship (AM) and rolling grants to Imperial College London and MSSL/UCL. 1

32 References Arridge, C. S., et al. (00), Modeling the size and shape of Saturn s magnetopause with variable dynamic pressure, J. Geophys. Res., 1, A. Axford, W. I., and C. O. Hines (), A unifying theory of high-latitude geophysical phenomena and geomagnetic storms, Can. J. Phys.,, 1,. Berchem, J., and C. T. Russell (1), The thickness of the magnetopause current layer ISEE 1 and observations, J. Geophys. Res.,,,. 1 1 Blanc, M., et al. (00), Magnetospheric and plasma science with Cassini-Huygens, Space Sci. Rev.,, Bunce, E. J., et al. (00), Origin of Saturn s aurora: Simultaneous observations by Cassini and the Hubble Space Telescope, J. Geophys. Res.,, A Clarke, J. T., et al. (00), Morphological differences between Saturn s ultraviolet aurorae and those of Earth and Jupiter, Nature,, Clarke, K. E., et al. (00), Cassini observations of planetary-period oscillations of Saturn s magnetopause, Geophys. Res. Lett.,, L. Cowley, S. W. H., et al. (00), A simple quantitative model of plasma flows and currents in Saturn s polar ionosphere, J. Geophys. Res.,, A01.

33 De Keyser, J., et al. (00), Magnetopause and boundary layer, Space Sci. Rev.,, 1. Dougherty, M. K., et al. (00), The Cassini magnetic field investigation, Space Sci. Rev.,, 1. Dougherty, M. K., et al. (00), Cassini magnetometer observations during Saturn orbit insertion, Science, 0, 1,. 1 Dungey, J.W. (1), Electrodynamics of the outer atmosphere, in Proceedings of the Ionosphere Conference, p., The Physical Society of London Eastman, T. E., and E. J. Hones Jr. (1), Characteristics of the magnetospheric boundary layer and magnetopause layer as observed by Imp, J. Geophys. Res.,,, Esposito, L. W., et al. (00), Ultraviolet imaging spectroscopy shows an active Saturnian system, Science, 0, 1, Fairfield, D. H., et al. (000), Geotail observations of the Kelvin-Helmholtz instability at the equatorial magnetotail boundary for parallel northward fields, J. Geophys. Res.,, 1,1.

34 Freeman, J. W., Jr., et al. (1), Plasma flow directions at the magnetopause on January 1 and 1 1, J. Geophys. Res.,,,1. Fujimoto, M., and T. Terasawa (1), Anomalous ion mixing within an MHD scale Kelvin-Helmholtz vortex, J. Geophys. Res.,,,01. Fukazawa, K., et al. (00a), Magnetospheric convection at Saturn as a function of IMF B z, Geophys. Res. Lett.,, L01. Fukazawa, K., et al. (00b), Vortex-associated reconnection for northward IMF in the Kronian magnetosphere, Geophys. Res. Lett.,, L Galopeau, P. H. M., et al. (1), Source location of Saturn s kilometric radiation: The Kelvin-Helmholtz instability hypothesis, J. Geophys. Res., 0,, Gérard, J.-C., et al. (00), Altitude of Saturn s aurora and its implications for the characteristic energy of precipitating electrons, Geophys. Res. Lett.,, L Goertz, C. K. (1), Detached plasma in Saturn s front side magnetosphere, Geophys. Res. Lett.,,. 1 Haaland, S. E., et al. (00), Four-spacecraft determination of magnetopause orientation, motion and thickness: comparison with results from single-spacecraft methods, Ann. Geophys.,, 1,.

35 Hasegawa, H., et al. (00), Transport of solar wind into Earth s magnetosphere through rolled-up Kelvin-Helmholtz vortices, Nature, 0,. Kivelson, M. G., and Z. Y. Pu (1), The Kelvin-Helmholtz instability on the magnetopause, Planet. Space Sci.,, 1,. Knight, S. (1), Parallel electric fields, Planet. Space Sci., 1, 1. Krimigis, S. M., et al. (00), Magnetospheric imaging instrument (MIMI) on the Cassini mission to Saturn/Titan, Space Sci. Rev.,,. 1 1 Lee, L. C., et al. (), Kelvin-Helmholtz instability in the magnetopause-boundary layer region, J. Geophys. Res.,, Lepping, R. P., et al. (), Surface waves on Saturn s magnetopause, Nature,, Lui, A. T. Y., et al. (1), Auroral bright spots on the dayside oval, J. Geophys. Res.,,, Masters, A., et al. (00), Surface waves on Saturn s dawn flank magnetopause driven by the Kelvin-Helmholtz instability, Planet. Space Sci., doi:.1/j.pss McAndrews, H. J., et al. (00), Evidence for reconnection at Saturn s magnetopause, J. Geophys. Res.,, A0.

36 McAndrews, H. J., et al. (00), Plasma in Saturn s nightside magnetosphere and the implications for global circulation, Planet. Space Sci., doi:.1/j.pss Mitchell, D. G.,et al. (00), Ion conics and electron beams associated with auroral processes on Saturn, J. Geophys. Res.,, A01. Miura, A. (1), Simulation of Kelvin-Helmholtz instability at the magnetospheric boundary, J. Geophys. Res.,,,1. 1 Nikutowski, B., et al. (00), Equator-S observation of reconnection coupled to surface waves, Adv. Space Res.,, 1, Nykyri, K., et al. (00), Cluster observations of reconnection due to Kelvin- Helmholtz instability at the dawnside magnetospheric flank, Ann. Geophys.,,, Nakamura, T. K. M., and M. Fujimoto (00), Magnetic reconnection within rolled- up MHD-scale Kelvin-Helmholtz vortices: Two-fluid simulations including finite electron inertial effects, Geophys. Res. Lett.,, L Ogilvie, K. W., and R. J. Fitzenreiter (1), The Kelvin-Helmholtz instability at the magnetopause and inner boundary layer surface, J. Geophys. Res.,, 1,.

37 Otto, A., and D. H. Fairfield (000), Kelvin-Helmholtz instability at the magnetotail boundary: MHD simulations and comparison with Geotail observations, J. Geophys. Res.,, 1,1. Owen, C., et al. (00), Cluster observations of surface waves on the dawn flank magnetopause, Ann. Geophys.,, 1. Phan, T. D., et al. (1), The magnetosheath region adjacent to the dayside magnetopause: AMPTE/IRM observations, J. Geophys. Res.,,. 1 Pu, Z. Y., and M. G. Kivelson (1), Kelvin-Helmholtz instability at the magnetopause: energy flux into the magnetosphere, J. Geophys. Res.,, Pu, Z. Y., and M. G. Kivelson (1), Kelvin-Helmholtz instability and MHD surface waves on Saturn s magnetopause, Chin. J. Space Sci.,, Ray, L. C., et al. (00), Current-voltage relation of a centrifugally confined plasma, J. Geophys. Res.,, A Richardson, J. D. (1), Thermal ions at Saturn Plasma parameters and implications, J. Geophys. Res., 1, 1,1. Russell, C. T., and R. C. Elphic (1), Initial ISEE magnetometer results Magnetopause observations, Space Sci. Rev.,, 1.

38 Saur, J., et al. (00), Anti-planetward auroral electron beams at Saturn, Nature,,. Sckopke, N., et al. (), Structure of the low-latitude boundary layer, J. Geophys. Res.,,,0. Southwood, D. J. (1), The hydromagnetic stability of the magnetospheric boundary, Planet. Space Sci., 1,. Takagi, K., et al. (00), Kelvin-Helmholtz instability in a magnetotail flank-like geometry: Three-dimensional MHD simulations, J. Geophys. Res., 1, A Talboys, D. L., et al. (00), Characterization of auroral current systems in Saturn s magnetosphere: High-latitude Cassini observations, J. Geophys. Res.,, A Tsurutani, B. T., et al. (1), Lion roars and nonoscillatory drift mirror waves in the magnetosheath, J. Geophys. Res.,,, Violante, L., et al. (1), Observations of mirror waves and plasma depletion layer upstream of Saturn s magnetopause, J. Geophys. Res., 0, 1,0. 1 Vo, H. B., and J. S. Murphree (1), A study of dayside auroral bright spots seen by the Viking aurorl imager, J. Geophys. Res., 0,,.

39 Young, D. T., et al. (00), Cassini plasma spectrometer investigation, Space Sci. Rev.,, 1. Zwan, B. J., and R. A. Wolf (1), Depletion of solar wind plasma near a planetary boundary, J. Geophys. Res., 1, 1,.

40 Figure captions Figure 1. Schematic illustrating the growth of the K-H instability at Saturn s dawn flank magnetospheric boundaries. The equatorial cross-section of the magnetosphere as viewed from the North is shown. The solar wind and magnetospheric plasma flows and the structure of the magnetospheric boundary region are indicated. 1 1 Figure. Trajectory of the Cassini spacecraft between 1 and 1 December 00 projected onto the x-y plane of the KSM coordinate system. Black dots on the trajectory indicate the position of the spacecraft at the beginning of each day. The cross-section of the Arridge et al. [00] magnetopause model scaled to intersect the final magnetopause crossing position is shown, as well as the position of the vortex encounter at the edge of the boundary layer Figure. MAG, ELS, and LEMMS observations for a h interval encompassing the boundary crossings of the Rev B inbound pass. (a) KSM components of the magnetic field. (b) Magnetic field magnitude. (c) Magnetic field elevation angle. (d) Electron number density (red) and temperature (blue) derived from ELS anode. (e) Timeenergy spectrogram of electron count-rate averaged over all ELS anodes. (f) Timeenergy spectrogram of electron intensity from LEMMS. (g) Time-pitch angle spectrogram of electron intensity from LEMMS. Spacecraft trajectory information is shown below the bottom panel. Shaded intervals correspond to when the spacecraft was in the magnetosheath. The pair of vertical dashed lines indicate the start and end of the vortex encounter. 0

41 1 1 Figure. MAG, ELS, IMS, and LEMMS observations for a h interval encompassing the spacecraft encounter with a vortex stucture on the edge of the boundary layer. (a) KSM components of the magnetic field. (b) Magnetic field magnitude. (c) Magnetic field elevation angle. (d) min resolution magnetic field perturbation vectors projected onto the M-N plane of the boundary normal coordinate system. (e) Electron number density (red) and temperature (blue) derived from ELS anode. (f) Time-energy spectrogram of electron count-rate averaged over all ELS anodes. (g) Time-energy spectrogram of ion count-rate from IMS detector. (h) Time-energy spectrogram of electron intensity from LEMMS. (i) Time-pitch angle spectrogram of electron intensity from LEMMS. Spacecraft trajectory information is shown below the bottom panel. The four sub-intervals are numbered, and the transitions between sub-intervals are indicated by changes in the background shading or vertical lines in each panel Figure. Schematic illustrating the structure of the vortex encountered by Cassini and passage of the spacecraft through it. A cross-section of the vortex in the M-N plane of boundary normal coordinates as viewed from the North is shown. The path of the spacecraft corresponds to the rest frame of the vortex. The crossed circles labeled B 1 and B represent the nominal orientation of the magnetic field in the magnetopshere proper and boundary layer respectively, which are both directed into the page. The different regions are indicated and the black arrows represent the anticipated directions of plasma flow. The gray arrow gives the sense of plasma circulation within the structure. The regions encountered by Cassini during the four sub-intervals are also indicated. 1

42 Figure. Schematic illustrating the local current system produced as a result of the formation of the K-H vortex. The vortex structure, magnetic field lines, and electric currents are shown as viewed from within the equatorial plane. The vortex is given as as a dark-gray oval, magnetic field lines in the foreground (background) are given as solid (dashed) black lines with arrow heads, and currents are given as light-gray arrows. 1 1 Figure. Schematic illustrating the global current system produced as a result of the formation of the K-H vortex. The vortex is given as a dark-gray oval, electric currents are given as black lines with arrow heads, Saturn s ionosphere is given as a dashed gray circle around the planet, and the hatched regions indicate the sites of possible field-aligned potentials that could lead to electron acceleration into and out of the ionosphere. 1

43 Figure 1.

44 Figure.

45 Figure.

46 Figure.

47 Figure.

48 Figure.