Cluster and SuperDARN observations during a positive B y period

1. Cluster and SuperDARN observations during a positive By period D. Ambrosino, E. Amata, M.F. Marcucci, I. Coco Istituto di Fisica dello Spazio Interplanetario, INAF Via del fosso del cavaliere, 100, 00133 Roma, Italy Introduction On January 2, 2003, around 23:00-24:00 UT Cluster crossed the northern dusk high latitude magnetopause on an outbound orbit. The IMF, observed by ACE, and the magnetosheath magnetic field, observed by GEOTAIL and by Cluster SC2, first changed from being predominantly northward to being By dominated and slightly southward; later on the IMF rotated slowly until it became purely northward. Before crossing the MP, Cluster SC3 observed a boundary layer of dawnward flowing plasma; during the same period, SC1 and SC4 stayed mainly in the magnetosheath and crossed the MP several times observing a similar BL of dawnward flowing plasma. Cluster observations can be explained in the framework of high latitude dusk reconnection at the cusp. Simultaneously, SuperDARN observations show that the high latitude convection at noon is characterized by an extended westward flow, when By > 0 and Bz < 0, and by sunward and westward flows when By > 0 and Bz > 0. We analyze the reconnection signatures at Cluster and study SuperDARN and Cluster data to describe the global reconfiguration of reconnection at the magnetopause driven by the IMF turnings.

2. Fig. 1 Fig. 1 showsACE and WIND positions in interplanetary space and GEOTAILlocation in the dawn flank magnetosheath for the event under study. 02/01/2003 x z z y y x GSM Fig. 2 SC1 SC2 SC3 SC4 Fig. 2 shows the Cluster outbound orbit during the crossing of the northern high latitude dusk magnetopause tailward of the cusp. The spacecraft formed a regular tetrahedron with 1 RE distance. SC3 was deeper in the magnetosphere; SC1, SC2 and SC4 were duskward with respect to SC3; SC2 had the most sunward position.

3. a b c d e f ACE WIND GEOTAIL SC2 SC2 MSH Fig. 3 IMF clock angle Fig. 3 displays solar wind density, magnetic field components and clock angle between 2003/01/02 23:00 and 2003/01/03 00:24 UT. The ACE and WIND data have been shifted to match the SC2 clock angle in the magnetosheath. At about 23:20 UT the IMF turneds rapidly from a northward to a duskward orientation, afterwards it slowly returned northward. Shaded boxes a through f highlight 2-min intervals further discussed hereafter (see Fig. 4).

4. Fig.4 shows six 2-min northern hemisphere SuperDARN convection maps corresponding to the intervals highlighted in Fig. 3. The maps are based on data from the following radars: Hankasalmi, Thykkvibaer, Stokkseyri, Goose Bay, Kapuskasing, Saskatoon, Kodiak, Prince George. • 23:10 - 23:12 UT. The polar cap convection is sunward and slightly westward. This is also observed during the previous 12 minutes and is compatible with reconnection tailward of the northern cusp. During this period the IMF z and y components are both positive. • 23:18 – 23:20 UT. Situation similar to panel a. However, the sunward/westward flows are less intense. At this time B, as observed at Cluster, becomes duskward and slightly southward. • 23:24 – 23:26 UT. The convection pattern displays a clockwise cell predominantly in the post noon sector with noon flows directed antisunward as expected for a By dominated IMF. The convection reconfiguration time is ~ 6 min. • 23:36 – 23:38 UT. The westward flows seem to enhance. Note that between 23:33 and 23:38 UT Bz becomes zero and the IMF is purely duskward. • 00:10 – 00:12 UT. The IMF is again northward with By > 0: correspondingly the flow rotates and becomes sunward/westward directed again. • 00:20 – 00:22 UT. Two reversed cells typical for a purely northward IMF are observed. This is confirmed by DMSP observations (see Fig. 5).

5. b a c d f e Fig. 4

6. Fig. 6a. Energy spectrograms and moments for SC3. BL (with dawnward flow) was entered at about 23:23 UT. MP was crossed at about 00:07 UT. Cluster 3 M P Fig. 6a

7. Fig. 6b. Same as Fig. 6a for SC4.The Boundary Layer (dawnward flow) was entered at about 23:20 UT. The magnetopause was crossed at about 23:40 UT (while SC3 exited later on at about 00:07 UT). Cluster 4 M P Fig. 6b

8. Fig. 6c. Same as Fig. 6a for SC1. BL was entered at 23:22 UT. MP was crossed at 23:38 UT. From 23:23 to about 23:26 UT (shaded area in the V panel) all SC (see also 6a and 6b) observed a duskward, tailward and northward flow at a lower velocity. The BL dawnward flow could be the effect of reconnection due to a duskward magnetosheath field. In the following we will show that this is indeed the case. Cluster 1 M P Fig. 6c

9. Changes in B and V across a reconnected MP must satisfy the Walén relation (Hudson, P.D., Planet. Space Sci., 1970, Paschmann, G. et al., Nature, 1979): Walén test The plus (minus) sign is for flow parallel (antiparallel) to B, i.e. , in our case, for observations sunward (tailward) of the reconnection site. Fig. 7 shows the Walén test result relative to the SC3 MP crossing of 00:07 UT, marked in Fig. 6a by a horizontal line between vertical arrows. The quality of the test is indicated by the red vector. Its length is the ratio between Vobs observed across the MP and the theoretical value; the angle between its direction and the horizontal reference line is the angle between Vobs and Vth. Thus, ideally the red vector should be of length one and parallel (antiparallel) to the positive horizontal axis for sunward (tailward) flow. Figs. 8a and 8b similarly show the Walén test results for two SC1 MP crossings, at 23:27 and 23:36 UT, marked in Fig. 6c by horizontal lines between vertical arrows. 23:27 UT 23:36 UT 00:07 UT Fig. 8b Fig. 8a Fig. 7

10. Kinetic signatures of reconnection are the so called D-shaped distribution functions of transmitted magnetosheath ions in the MP/BL. D-shaped distribution functions are expected in the MP/BL with a low energy cutoff at a velocity along the magnetic field equal to the parallel component of the deHoffmann-Teller (dHT) velocity (Cowley, 1982). During this event, D-shaped distribution functions were observed almost continuously in the BL by SC1, SC2 and SC3. For one case (not shown herein) we verified that the cutoff corresponds to the parallel component of the MP dHT velocity. • Fig. 9 displays three D-shaped distributions in the V// - V^ plane. We notice that: • their flow is parallel to the BL field, • their cutoff ranges from a few km/s to about 100 km/s, • they are mixed with the magnetospheric plasma, • at times a second antiparallel population appears (see e.g. the SC4 observations of 23:22:10 UT). The second antiparallel population can be interpreted as the transmitted m’sheath population reflected back from the ionosphere. When such a population is present (see SC4 distribution function at 23:22:10) the computed velocity results to be very low. SC1 23:21:59 UT SC4 23:22:10 UT SC3 23:23:11 UT Fig. 9 From the analysis of the CLUSTER data we conclude that the dawnward jets observed in the BL are due to reconnection.

11. Fig. 10 Bewteen 23:18 and 23:23 all the Cluster satellites observe D-shaped distributions.The transmitted populations are directed dawnward, sunward and southward (note that computed velocity for SC1 and SC4 in the last part of this interval is low since a second reflected population is present). From 23:23 to about 23:26 (shaded area) all the spacecraft observe a plasma which is moving duskward, tailward and northward at a much lower velocity. After 23:26, a reversal in the plasma flow direction occurs: the flow being directed dawnward, but antisunward and northward. Since at about 23:20 the IMF changed from northward to duskward and slightly southward we interpret these observation as reconnection first occurring tailward of Cluster, then ceasing and initiating at high latitude duskward of the three spcecraft. In Fig. 10 we show a zoom of the Cluster velocities and magnetic fields between 23:10 and 23:40 UT. The 23:20 – 23:30 UT period is highlighted. Note: we checked that the Vx negative jets seen by SC1 at 23:28:30 and at 23:29 UT are not due to reconnection but correspond to partial MP crossings. SC3 SC3 SC1

12. 23:20 – 23:24 UT maps. The convection reconfiguration begins at low latitudes as the antisunward flow begins to appear at about 75° MLAT 11-12 MLT. At this time the Cluster footprints are slightly above 80° MLAT and 12 MLT, where the convection is still sunward. This is in agreement with the sunward flows observed by Cluster at the magnetopause. 23:24-23:26 UT map. The Cluster footprint is in the center of the developing westward cell in a region of low convection. We recall that at this time reconnection seems to stop at the MP close to Cluster. As the convection reconfiguration occurred between 23:22 and 23:26 UT (cf. Fig. 4), in the following we concentrate on the 23:20-23:30 UT period. Fig. 10 shows Cluster and FAST footprints drawn on SuperDARN maps. 23:26-23:28 UT map. The Cluster footprint is embedded in the westward and antisunward flow, again in agreement with observations made by Cluster at the MP. As regards FAST, we notice that its footprint moves from dawn to dusk and towards lower latitudes in a region where SuperDARN observes tailward-westward convection. a b c d Fig. 11 Fast Cluster

13. Fig. 12 Fig.5 Fig.12 shows the ACE IMF components for 23:10 - 23:30 UT and the FAST proton energy spectrum for 23:20 – 23:30 UT. From 23:24 UT onwards proton energy increases as latitude decreases, suggesting that a reconnection site is being approached. Actually, we showed that reconnection starts duskward of the cusp between 20:20 and 20:24 UT (cf. Fig. 11). Several energy decreases occurring on time scales of 10-20 S need to be further studied. At 00:20 UT, when IMF Bz > 0, the SuperDARN maps show a sunward convection over the polar cap. Fig. 5 shows the DMSP F13 trajectory (east to west) and the polar plot of Cross-Track Plasma Drift. We clearly see a sunward flow around noon.

14. WARNING: in Fig. 15 dawn is on the left and the northern polar cap on the top; in Fig. 16 dawn is on the right and low latitude is on the top. Summary Fig. 13c Fig. 13a Fig. 13b 23:26:00 – 23:28:00 UT 23:22:00 – 23:24:00 UT 23:20:00 – 23:22:00 UT Fig. 14c Fig. 14a Fig. 14b Cluster “X line” Reconnection jets and correspondent ionospheric flows. a) Vy < 0, Vx > 0 (t < 23:22 UT). Reconnection tailward of the cusp. b) Vy» 0 (23:22 UT < t < 23:26 UT). No reconnection signatures at Cluster. “By reconnection” is approching from lower latitudes. c) Vy < 0, Vx < 0 (23:26 UT < t) By > 0,Bz < 0reconnection duskward of the cusp.

15. Conclusions Between 23:00 UT, 2 January, and 01:00 UT, 3 January, 2006, the IMF underwent several orientation changes. The SuperDARN convection maps, the DMSP and FAST observations and the observations made by Cluster (in its outbound orbit) at the MP, in the MS and in the BL all agree on a global reconfiguration of MP reconnection and ionospheric convection driven by IMF changes. The period between 23:20 and 23:30 UT was analysed in detail. At about 23:20 UT the IMF turned rapidly from a northward to a duskward and slightly southward orientation. At the time of the IMF turning reconnection tailward of the cusp was ongoing. The 23:20 UT IMF rotation drove a reconfiguration of reconnection at the MP (Cluster) and of ionospheric convection (SuperDARN). At the IMF rotation, the cusp dusk side reconnection started, while reconnection tailward of the cusp was still going on (23:20-23:24 UT; cf. Figs. 11a,b). Within 2 min tailward reconnection stopped. From 23:26 UT onwards the cusp dusk side reconnection extended to higher latitudes reaching Cluster. FAST observed on average an energy dispersion (energy increased as latitude decreased) suggesting that a reconnection site was being approached. At the FAST footprint (9-13 MLT and 80°-75° magnetic latitude) SuperDARN observed anti-sunward flows. Such observations are reconciled under the hypothesis that the flows came from a By dominated reconnection line duskward of the cusp.