TC 1 observations of a flux rope: Generation by multiple X line reconnection

Transcription

1 JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 116,, doi: /2010ja015986, 2011 TC 1 observations of a flux rope: Generation by multiple X line reconnection L. Trenchi, 1 M. F. Marcucci, 1 H. Rème, 2 C. M. Carr, 3 and J. B. Cao 4 Received 2 August 2010; revised 24 January 2011; accepted 15 February 2011; published 10 May [1] On 29 December 2004 the Double Star TC 1 satellite most probably stays close to a reconnection line for about 20 min. The direction of the X line can be evaluated according to the component merging model. During the time of observations, TC 1 also detects a flux transfer event (FTE). It is found that the FTE axis is parallel to such an X line. This observation is more in favor of the single X line and the multiple X lines FTE models, rather than the original Russell and Elphic model. Moreover, particular features of the ion distribution functions and the plasma behavior near the FTE seem to be more consistent with a multiple X line generation mechanism. The period under study is characterized by a large and negative Earth s dipole tilt, and the reconnection line close to TC 1 is found to follow the magnetic equator, passing northward of the subsolar/stagnation point. These facts lead to speculate that the FTE is generated when a secondary X line, probably passing through the subsolar/stagnation point and with approximately the same orientation of the X line near TC 1, is activated southward of TC 1. Citation: Trenchi, L., M. F. Marcucci, H. Rème, C. M. Carr, and J. B. Cao (2011), TC 1 observations of a flux rope: Generation by multiple X line reconnection, J. Geophys. Res., 116,, doi: /2010ja Introduction [2] The term flux transfer event (FTE) has been used for the first time by Russell and Elphic [1978] to define the transient magnetic field perturbations observed by ISEE 1 and 2 at the low latitude magnetopause. These perturbations consist principally of bipolar signature in the component of the magnetic field normal to the magnetopause. The model proposed by Russell and Elphic [1978] predicts a pair of elbow shaped flux ropes of reconnected field lines generated by intermittent and localized reconnection. These flux ropes extend outward into the solar wind, cross the magnetopause through approximately circular holes and, inward, reach the ionosphere. The draping of the surrounding fields around these flux ropes produces the bipolar signature. Lee and Fu [1985] proposed another magnetic flux transfer model based on multiple X lines reconnection. The simultaneous reconnection of the magnetospheric and the magnetosheath fields at different X lines generates flux ropes which have the axis roughly aligned to the X lines. The bipolar signature is again expected, but in this model the flux ropes could have considerable extension along their axis. A third model [Scholer, 1988; Southwood et al., 1988] is based on bursty reconnection at a single X line. In this case an enhancement and a subsequent reduction of the reconnection rate at the X line generates bubble like structures at the magnetopause which 1 INAF Istituto di Fisica dello Spazio Interplanetario, Rome, Italy. 2 CESR, Toulouse, France. 3 Imperial College London, London, UK. 4 Beijing University of Aeronautics and Astronautics, Beijing, China. Copyright 2011 by the American Geophysical Union /11/2010JA propagate along the magnetopause and produce the bipolar signature. Flow shears across the magnetopause and/or nonantiparallel fields produce structures with twisted field lines. Also these structures have the axis roughly aligned to the X line, but these have not the helical internal field lines. [3] Many studies based on the magnetic field, electric field, plasma and particles observations confirmed that FTEs are generated at the magnetopause by magnetic reconnection related phenomena [Berchem and Russell, 1984; Zong et al., 2003; Wang et al., 2006]. However, the details of the generation mechanism responsible for FTEs remain unclear. For example, Kawano and Russell [2005] performed a statistical analysis based on 634 FTEs simultaneously observed by ISEE 1 and ISEE2. Their results show that the flux ropes longitudinal scale is shorter than the latitudinal scale, as for the flux ropes generated by the Russell and Elphic [1978] model. On the other hand, Lockwood and Hapgood [1998], studying the ion spectrogram for an FTE observed by AMPTE, concluded that their observations were consistent with the single X line FTE model. More recently, several studies gave indication that the Lee and Fu [1985] multiple X lines model can account for FTEs generation [Zong et al., 2005; Hasegawa et al., 2006; Zhang et al., 2008; Hasegawa et al., 2010]. [4] The detailed analysis of the flux ropes structure in relation with the reconnection geometry can give useful information about their generation mechanism. On 29 December 2004 the Double Star TC 1 satellite, detects the passage of an FTE. Moreover, TC 1 observes several reconnection flow reversals, indicating that the satellite is close to a reconnection line. The direction of the X line, as predicted by the component merging model, is found to be parallel to the FTE axis and tangent to the magnetopause. In this paper, the possible 1of11

2 Figure 1. The positions of TC 1, SC1, and Geotail at the time of observations in (a) the X GSM Z GSM plane and (b) Y GSM Z GSM plane. generation mechanism of the FTE and the role played by the Earth s dipole tilt, very large at the time of observations, are discussed. 2. Observations [5] The present study uses the observations made with the HIA [Rème et al., 2005] and the FGM [Carr et al., 2005] experiments on board TC 1 in the UT interval. The interplanetary magnetic field (IMF) and the solar wind conditions are provided by Geotail, orbiting in the solar wind close to the dawn magnetopause, by ACE located at the Lagrangian point L1 and by Cluster SC1, which stays in the southern magnetosheath at the same local time as TC 1. TC 1, SC1 and Geotail positions in the GSM reference are (5, 10, 0), (9,11, 10) and (24, 14, 0), respectively (see Figure 1). [6] In Figures 2a 2d, the magnetic fields measured by SC1, Geotail, ACE and TC 1 are reported. The plasma density and velocity measured by ACE are reported in Figures 2e and 2f. The magnetic field clock angles (tan 1 (B y /B z )) measured by TC 1, SC1, Geotail and ACE are plotted in Figure 3a. In these plots, the data measured by ACE and Geotail have been forward shifted by 56 and 3 min, respectively, to take into account the convection time. The ion density and velocity computed onboard TC 1 from the three dimensional distribution functions at four seconds time resolution are plotted in Figures 3b and 3c. The magnetic field measured by TC 1, averaged at the same time resolution, is reported in Figure 3d. Figures 3e and 3f will be described later on. The GSM reference frame is used in Figures 2 and 3. The grey shadings highlight the time interval when TC 1 observes the magnetic structure studied in the present paper, analyzed in detail in the next paragraphs. [7] Looking at Figures 2a 2d and Figure 3a, it can be noted that during the event the IMF is directed southward and dawnward, apart from the short northward turning at about The solar wind density and velocity are quite stable during the event and the average values are about 10 cm 3 and 430 km/s, respectively (see Figures 2e and 2f). [8] TC 1 stays mostly in the magnetosphere before 0757 and exits definitively in the magnetosheath at During the interval, TC 1 repeatedly crosses the magnetopause (MP) and the boundary layer (BL), a region inward of the MP where the density and the field magnitude assume intermediate values between the magnetosheath and magnetospheric levels. It can be noticed that accelerated plasma flows are often present both at the MP and in the BL layer. The accelerated plasma flows move, with respect to the magnetosheath, in two opposite directions: southward, dawnward and sunward (e.g., the jets at 0810:30 and 0814) and northward, duskward and tailward (e.g., the jets at 0818 and 0838). Moreover, a particular magnetic field structure is detected in the interval. Such structure is evidenced by the grey shadings in Figures 2 and 3. The magnetic field in this time period seems to be different from both the magnetosheath and the magnetospheric fields since SC1, Geotail and ACE do not observe any similar field rotation. In sections 2.1 and 2.2, it will be shown that the accelerated flows are caused by magnetic reconnection and that TC 1 is close to the reconnection line. In sections 2.3 and 2.4, instead, a detailed study of the magnetic structure will be presented which gives indication on how such a structure is generated at the magnetopause Reconnection Jets [9] In order to check whether the accelerated plasma jets are due to reconnection, the Walén test is performed, that is, it is checked whether the observed plasma acceleration is due to the magnetic tension force and, therefore, whether the following relation holds [Paschmann et al., 1986 and references therein]: V J V 1 ¼½ð1 1 Þ= 0 1 Š 1=2 ½B J ð1 J Þ= ð1 1 Þ B 1 Š ð1þ Here B, V, r and a are the magnetic field vector, the plasma velocity, the mass density and the anisotropy factor, respectively. The subscript 1 refers to the magnetosheath adjacent to the magnetopause and, in this study, it indicates 2of11

3 Figure 2. (a d) The magnetic fields measured by SC1, Geotail, ACE, and TC 1 are reported. Two different scales are used for the satellites located in the solar wind, Geotail, and ACE (left scale) and in the magnetosheath, SC1, and TC 1 (right scale). (e f) The plasma density and the velocity measured by ACE are reported. The GSM reference frame is used. The data measured by ACE and Geotail have been forward shifted by 56 min and 3 min, respectively, to consider the convection time. The grey shadings highlight the time interval when TC 1 observes the FTE. averages computed in the interval, taken as the magnetosheath reference and highlighted by the yellow shading in the Figure 3; the subscript J indicates values on the magnetospheric side of the magnetopause. The positive (negative) sign applies to observations north (south) of the reconnection line. In order to verify equation (1), the ratio R W between the strengths of the observed velocity jump (V J V 1 ) and the expected one (right hand side of the Walén relation), along with the angle Q W between these two vectors, have been computed for each data point in the interval under study. R W and Q W are reported in Figures 3e and 3f. The Walén relation is perfectly satisfied when R W is 1 and Q W is 0 (180 ) for reconnection jets north (south) of the reconnection point. Considering the applications of the Walén test reported in the literature on observations at the magnetopause, it is found that usually the observed and theoretical velocity changes across the magnetopause are well aligned (Q W is 0 or 180 ), while R W is often lower than unity. [10] Several possible reasons have been proposed to explain such not perfect agreement [Paschmann and Sonnerup, 2008 and references therein]. Here, following Paschmann et al. [2005] and Phan et al. [1996], we consider as due to reconnection the flows which satisfy the conditions R W > 0.5, 0 < Q W < 30 (or 150 < Q W < 180 ) and n > 1(cm 3 ), this last threshold assuring that the plasma of the accelerated flows is of magnetosheath origin. In Figure 3 the pink and blue shadings evidence the plasma jets, at the magnetopause or in the boundary layer, for which the aforementioned conditions are satisfied. The pink shading is used when 0 < Q W < 30 and it indicates reconnection jets north of the reconnection line, analogously, the blue shading indicates reconnection jets south of the reconnection line. It can be concluded that for the majority of the accelerated flows the reconnection conditions are satisfied and, consistently with the expectations, for the northward (southward) jets the Walén test is fulfilled with a plus (minus) sign, being the satellite north (south) of the reconnection line. Moreover, the passage 3of11

4 Figure 3. (a) The magnetic field clock angles measured by SC1, Geotail, ACE, and TC 1. (b) Ion density, (c) velocity, (d) the magnetic field, (e) R W, and (f) Q W. R W and Q W are the ratio and the angle between the variation of velocity predicted by the Walén relation across the magnetopause and the observed one, respectively. The GSM reference system is used. from northward to southward reconnection jets, and vice versa, could be an indication of the passage of the satellite from the northern side to the southern side, and vice versa, of a reconnection X line Ion Distribution Functions [11] There are a number of typical kinetic signatures in the distribution functions of particles streaming along reconnected field lines which have been predicted by kinematic considerations [Cowley, 1982, 1995] and later confirmed by observations [Fuselier, 1995 and references therein]. In particular, Bavassano Cattaneo et al. [2006] show that the reconnection site was near the Cluster spacecraft during a lobe reconnection event throughout a detailed study of the ion distribution functions. Similarly to these authors, we will show kinetic evidence of reconnection and discuss some aspect of the observations which are related to the issue of the vicinity of TC 1 to the reconnection site. [12] Cuts in the V par V per plane of some characteristic ion distribution functions are shown in Figure 4. The V par V per plane is defined by the two vectors: B, at the central time of the ion distribution measurement, and by the vector (V B) B, where V the ion bulk velocity vector. A magnetosheath distribution function, measured in the 0813: :27 interval, is shown in Figure 4a. In Figure 4b it is shown the distribution function measured during the 0813: :11 interval. During this interval, TC 1 is in the magnetosheath boundary layer (MSBL), the region outside the MP where the magnetic field initiates to rotate. TC 1 detects a heated incident magnetosheath population which has enlarged contours in the direction antiparallel to the magnetic field. This can be interpreted as due to the presence of a secondary population which flows along the convected magnetic field direction. Such population is displaced, with respect to the incident population, by twice the local Alfvén speed 2V A, along the convected magnetic field direction (V A = 300 km/s), as predicted for magnetosheath ions reflected off the MP within the current layer [Fuselier, 1995]. The fact that the reflected ions stream antiparallel to the MSBL field means that the spacecraft is located northward of the reconnection line. It can be noted that the BL observed before the magnetopause rotation at 0813 presents a reconnection jet flowing southward, which means that the spacecraft is located southward of the reconnection line. Therefore the reconnection line passes over the spacecraft within 36 s. [13] In Figures 4c and 4d two distributions which have the D shape as expected in case of reconnection are shown. These distribution functions are measured in the boundary layer during the intervals 0816: :50 and 0823: :50, respectively. The distribution of Figure 4c shows a 4of11

5 transmitted magnetosheath population which flows parallel to the magnetic field, being the satellite northward of the reconnection line. Conversely, in the distribution of Figure 4d the transmitted population flows antiparallel to the magnetic field and the satellite is located southward of the reconnection line. The low energy cutoff in the D shaped distributions should correspond to the de Hoffmann Teller velocity. The de Hoffmann Teller velocity (V HT ) relative to the distribution function of 0816:42 has been evaluated with the minimization of the residual electric field [Paschmann and Sonnerup, 2008 and references therein] across the MP of The results of the de Hoffmann Teller analysis for this crossing are shown in Figure 5a and the projection of V HT in the V par V per plane is marked by the black cross in Figure 4c. The distribution of 0823:42 has no adjacent MP crossing and we estimated the de Hoffmann Teller velocity through the Cowley and Owen [1989] formula V HT CO = V 1 ± V A ^b 1, with the plus (minus) sign for observations south (north) of the reconnection line and bearing in mind that the plasma parameters and the magnetic field in the magnetosheath do not change much during the event. The black dot in Figure 4d marks the projection of V HT CO derived with the plus sign, since TC 1 is located south of the reconnection line. For comparison, the V HT CO derived for observation north of the reconnection line is indicated with a black dot in Figure 4c, as well. It can be seen that the two different estimates of de Hoffmann Teller velocities derived for the northward jet of 0816:42 (Figure 4c) are in good agreement. Moreover, the low energy cutoff of the distributions corresponds rather well to the de Hoffmann Teller velocity for both the northward and southward jet. It should also be noted that the velocity corresponding to the observed cut off in the 0823:42 distribution function (Figure 4d) is directed southward, which means that the reconnected magnetic field lines contract in that direction. This could be the case since M A is usually lower than 1 for this event [Gosling et al., 1991; Bavassano Cattaneo et al., 2006]. [14] In conclusion, the distribution function characteristics agree well with the prediction for reconnection. Furthermore, TC 1 is very probably near a reconnection line during the first part of event. However, it must be stressed that no inference on the length of such an X line can be obtained from the TC 1 single point observations. The distribution function in Figures 4e 4i will be discussed in section Observations in the X Line Reference Frame [15] The observations presented so far seem to indicate that TC 1 is near the reconnection line, therefore the orientation of such an X line can be computed according to the component merging model [Sonnerup, 1974; Gonzales and Mozer, 1974]. This model predicts that the X line is perpendicular to the vector B 2 B 1, where B 1 and B 2 are the magnetosheath and the magnetospheric field respectively. Afterward, it is possible to define a local reference frame with the ^n axis along the local magnetopause normal outward, the ^X L axis parallel to the X line (directed as the guide field) and the ^R C axis along the reconnecting field compo- 5of11 Figure 4. Cuts of the HIA ion distribution functions in the V par V per plane: (a) in the magnetosheath, (b and g) in the MSBL, (c f) in the BL, and (h and i) within the FTE. The black crosses represent the projection of the velocity obtained with the minimization of the residual electric field, V HT. The black dots are the projections of the de Hoffmann Teller velocity obtained with the Cowley and Owen formula, V HT CO. The bar with rounded ends corresponds to 2V A.

6 nent. The empirical magnetopause model of Fairfield [1971] is used to find ^n. The ^X L is given by ^X L ¼ ^n ð B 2 B 1 Þ j^n ðb 2 B 1 Þj where B 1 is the average of the magnetic field in the magnetosheath reference, while B 2 indicates the average of the magnetospheric magnetic field in the 0756: :30 interval. It is found that ^n = (0.71; 0.71; 0.01), ^X L = (0.58; 0.59; 0.56) and ^R C = (0.41; 0.39; 0.83), in GSM coordinates. [16] It is convenient to study the magnetic structure observed during the interval in the reference frame defined above. The velocity and the magnetic field observations in the new reference for the time interval are shown in Figure 6. It can be noted that the B n component is generally small, so that the ^n direction obtained through the Fairfield [1971] model can be considered as a good approximation of the real magnetopause normal. The magnetic field in the magnetosheath, after 0842, and in the magnetosphere, in the and intervals, is along the ^R C axis. Specifically, it has a positive B RC component in the magnetosheath and a negative B RC component in the magnetosphere. In fact, the magnetosheath and magnetosphere fields are almost antiparallel (the shear angle between B 1 and B 2 is 175 ) and the B XL component, that is the guide field, is very small. The magnetic structure, instead, is characterized by a large B XL and a B RC close to zero. Moreover, the B n component inside the structure shows a bipolar signature. In the central part of the structure, around 0830:40, where B n is zero, the field is exactly parallel to ^X L. [17] These signatures, schematically illustrated in Figure 7a, are consistent with the passage of TC 1 inside a flux rope (Figure 7b). The fact that B RC is small suggests that TC 1 is passing close to the center of the flux rope, as represented by the dot dashed path in Figure 7b. Indeed, the passage of the satellite in the magnetosheath (magnetospheric) side of the flux rope, would results in a positive (negative) B RC component. Therefore, the flux rope axis should be directed approximately along the magnetic field as measured close to the center of the flux rope, i.e., where B n is zero. Such axis is tangent to the magnetopause, aligned to the direction of the X line and directed as the guide field. In section 2.4, the orientation and motion of the magnetic structure will be determined more quantitatively. Figure 5. Results from the de Hoffmann Teller analysis for the MP crossings of (a) 08:16, of (b) 08:27, and of (c) 08:33. Scatterplots of the three components of the convection electric field versus that based on the V HT velocity in the spacecraft rest frame The Magnetic Structure Orientation [18] The least squares technique based on Faraday s law presented by Sonnerup and Hasegawa [2005] has been used to determine the axis orientation, motion and intrinsic electric field of the magnetic structure. This technique can be performed for plasma/magnetic field structures moving past a single observing spacecraft provided two conditions are fulfilled: (1) the structure is two dimensional (the derivatives along the axis are much smaller than those in the plane transverse to it); and (2) in the reference moving with the structure, the structure is time independent or nearly time independent. If this is the case, the Faraday s law requires the axial electric field to be constant in any frame of reference moving with such a structure (in the complete absence of dissipation, it must be zero). 6of11

7 Figure 6. Plasma and magnetic field observations in the X line reference frame. The grey shading highlights the magnetic structure. The format is the same as Figure 3. [19] The method has been applied following the steps described in section 3 of Sonnerup and Hasegawa [2005]. With the magnetic and electric fields fluctuations de =(E hdei) and db =(B hdbi), the covariance matrices M EB = hdedbi, M BE = hdbdei, M EE = hdedei, M BB = hdbdbi and the inverse matrix M BB have been evaluated. The axial direction of the structure, ^k, is found to be the eigenvector corresponding to the smallest eigenvalue of the symmetric matrix M 0 = M EB M BB M BE + M EE. A large ratio between the intermediate and the minimum eigenvalue is an indication that ^k is close to the true axial direction. In the application of the method, small fluctuations of the axial electric field about its average are considered as an indication that the observed structures satisfies well the two model assumptions. The velocity U 0 = M BB M BE k has been calculated and has been checked that the product ^k U 0, which is proportional the fluctuations of the axial electric field, is small when compared to U 0. Finally the proper velocity of the structure V 0 = k U 0 has been determined. In the case under study, the measured ion convection electric field E = V B has been used as a proxy for the total field, assuming all other terms in the generalized Ohm s law to give negligible, or constant, axial contribution. The ratio between the intermediate and the minimum eigenvalues of the matrix M 0 is about 13 and the eigenvector corresponding to the smallest eigenvalue, i.e., ^k, is (0.54; 0.59; 0.60) in GSM coordinates. It results ^k U 0 / U 0 < The velocity of the structure V 0 is found to be ( 34.6; 55.0; 85.2) km/s in GSM coordinates and the intrinsic axial electric field is E 0 = mv/m. The FTE axis direction ^k is along the direction inferred for the X line: the relative angle between ^k and ^X L is 2.9. The velocity V 0 is roughly antiparallel to ^R C (the relative angle is 172 ). Thus the satellite moves from the top to the bottom of the FTE (Figure 7b), in agreement with the observed positive negative B n signature. The length of the FTE cross section along the satellite path can be estimated by multiplying V 0 by the time extent during which TC 1 remains inside the flux rope and it is found to be approximately 3 R E Ion Distributions Nearby and Within the FTE [20] It can be noted that MP crossings occur just before and soon after the passage of the FTE. Specifically, a partial crossing from the BL to the MSBL occurs at 0827 and a partial crossing from the MSBL to the BL occurs at The distribution functions measured in the BL just before and soon after the passage of the FTE show the simultaneous presence of two dense populations flowing parallel and antiparallel to the magnetic field. As an example, the V par V per cuts of the distribution function measured in the time intervals 0826: :56 and 0833: :16 are shown in Figures 4e and 4f. The de Hoffmann Teller 7of11

8 Figure 7. (a) Magnetic signatures expected when the satellite passes close to the flux rope center along the ^R C direction. (b) Scheme of the flux rope cross section. The flux rope axis is aligned with the ^X L axis. analysis has been performed for the partial MP crossings occurring at about 0827 and 0833 and the results are shown in Figures 5b and 5c. A V HT velocity can be identified for each crossing. The black crosses in Figures 4e and 4f represent the projections of the V HT for the MP of 0827 and 0833, respectively. It can be noted that the population which flows antiparallel (parallel) to B in the distribution of 0826:48 (0833:08) is D shaped with the low energy cutoff at the southward (northward) V HT. Therefore, the characteristics of the antiparallel (parallel) population in Figure 4e (Figure 4f) are as expected in the case of magnetosheath plasma transmitted across the MP south (north) of the reconnection line. Moreover, the population antiparallel to B in Figure 4f is similar to the antiparallel population in Figure 4e and the B parallel population in Figure 4e resembles the parallel population in Figure 4f, although it is less dense. These observations are consistent with the TC 1 satellite being in between two X lines at the time of observations. In particular, the distribution functions examinated so far seem to be accounted for in the scheme in Figure 7b. TC 1 isin between two X lines from about 0827 to It is closer to the northern X line at the partial crossing of 0827, then it observes the passage of the magnetic structure and finally it detects the signatures of the southern reconnection line at the partial MP crossing of The results of the de Hoffmann Teller analysis for the MP crossings just before and after the FTE confirm the above interpretation. In the frame of reference of the FTE, the projections along ^R C of the V HT computed for the MP crossings of 0827 and 0833 are 84 km/s and 186 km/s, respectively. Therefore, the field lines before and after the structure are contracting toward the structure, in agreement with the proposed scheme. Moreover, the presence of two reconnection lines at the time of the FTE observation seems to be confirmed when the distribution function of 0828:33 is examined (Figure 4g). At this time TC 1 is in the MSBL and three distinct populations can be observed in the distribution function: (1) the incident magnetosheath population; (2) the magnetosheath population reflected at the MP southward of the northern reconnection line, parallel to B and displaced with respect to the incident population by 2V A ; and (3) the magnetosheath population reflected at the MP northward of the southern reconnection line, flowing antiparallel to B and displaced with respect to the incident population by 2V A (V A = 250 km/s). [21] It must be noted that the simultaneous presence of the two transmitted populations could also be interpreted as due to the passage past the spacecraft of a single X line, during the 8 second integration time. Nevertheless, this explanation seems to be less probable, since (1) the two transmitted populations are observed for at least three consecutive distributions and (2) the V HT for the MP crossing of 0827 (0833) gives an indication that TC 1 remains constantly southward (northward) of a northern (southern) reconnection line. Furthermore, regarding the possibility that the multiple populations are generated by the reflection of the transmitted magnetosheath population at the ionosphere, it seems that the characteristics of the distribution functions illustrated in Figures 4e and Figure 4f of are not compatible with such a picture. In particular, the reflected and the transmitted populations should be similar and should move at the same parallel velocity, but in the opposite direction. [22] Hereafter, two distribution function within the magnetic structure will be discussed. The V par V per cut of the distribution function measured in the time interval 0829: :22 is shown in Figure 4h). At this time, the magnetic field points along ^X L and has a positive B n component. Only two counter streaming populations can be observed in the distribution function; these populations have a higher temperature with respect to the ones observed in the adjacent MSBL, especially in the perpendicular direction. Going toward the center of the FTE, the two populations come closer and, in the internal part of the structure, only one population is present. As an example, the V par V per cut of the distribution function measured during the 0830: :35 interval is shown in Figure 4i. Such distribution is similar to the one observed in the magnetosheath (Figure 4a), but it has a lower value in the peak and it is slightly hotter in the direction parallel to the magnetic field. Actually, when compared to the adjacent MSBL, it has a higher perpendicular temperature, as already noted for the distribution in Figure 4h). In addition, considering higher energy particles, it is to be noted that all the distribution functions in the center and in the peripheral regions of the FTE show the presence of an almost isotropic magnetospheric population at velocities above 600 km/s, with a slight depletion for velocities antiparallel to the magnetic field, especially in the FTE core. 3. Discussion [23] The study of the plasma and magnetic field measurements provided by the instruments on board TC 1 gives useful hints on the configuration of the reconnection at the TC 1 position during the time of observations. In fact, 8of11

9 several passages from southward to northward jets and vice versa, together with the passage from the MSBL northward of the X line to the BL southward of the X line within 36 s, indicate that the spacecraft is near to the reconnection line. It is therefore meaningful to infer the orientation of such an X line according to the component merging model. In addiction, TC 1 also observes the passage of a magnetic structure characterized by the typical B n bipolar signature of the flux transfer events and the expected mixture of magnetosheath and magnetospheric plasma. The orientation of the axis of this magnetic structure has been evaluated using the technique developed by Sonnerup and Hasegawa [2005]. Comparing the results of the Sonnerup and Hasegawa [2005] method with the component model prediction for the orientation of the reconnection line, it is found that the magnetic structure axis is parallel to the direction of the X line. This observation gives important information about the generation mechanism of the magnetic structure. In fact, it indicates that this structure is most probably generated within the framework of the single X line or the multiple X line FTE models, being less probable that it is generated by patchy reconnection, as in the model originally proposed by Russell and Elphic [1978]. Regarding the single X line and multiple X lines models, it is not easy to discriminate between the two, but some facts seem to favor the latter. Precisely, in the BL just before and soon after the passage of the FTE, two counter streaming magnetosheath populations are observed, probably transmitted from two reconnection lines northward and southward of TC 1. Moreover, the de Hoffmann Teller analysis performed at the two MP partial crossing just before and after the FTE shows that the field lines are contracting toward the FTE. These signatures seem to suggest that two reconnection sites are active northward and southward of the magnetic structure and that the FTE is produced by the multiple X line model. Furthermore, the velocity and scale size of the observed structure (110 km/s and 3 R E ) seem to be on the lower (velocity) and upper (scale) ends of the statistical distributions of FTE reported from Cluster observations [Fear et al., 2007].This is more in favor of the multiple X line model, as the magnetohydrodynamic simulation of Ku and Sibeck [2000] shows that this model produces larger and slower structures than single X line model. A possible generation mechanism of the magnetic structure is discussed in section 3.1, taking into account the preceding considerations and the fact that the dipole tilt is very large at the time of the observations. Finally, in section 3.2 the core magnetic field enhancement is discussed Generation Mechanism [24] The X line located near TC 1, according to the orientation calculated in section 2.3, is tilted with respect to the Y GSM axis. Therefore, prolonging this X line from the TC 1 position, it intersects the noon meridian plane several Earth s radii northward of the subsolar point. We noted that, at the time of observations, the Earth s dipole tilt is 28.7, so that the magnetic equator is located northward of the subsolar point. Two recent statistical studies found a correlation between the X line position at the MP and the Earth dipole tilt. Trattner et al. [2007] analyzed 130 passes of Polar spacecraft through the northern cusp and estimated the position of the reconnection line by studying the time offlight characteristics of the ions precipitating in the cusp. On the other hand, analyzing 143 Double Star TC 1 reconnection events, Trenchi et al. [2008] found 23 crossings during which TC 1 was near the reconnection X line (the event under study is one of these 23 crossings). Both studies found a correlation between the position of the X lines and the dipole tilt: it seems that the X line follows the position of the magnetic equator northward (southward) of the subsolar point according to the negative (positive) tilt. TC 1 observes an FTE whose axis is parallel to the X line. The properties of the distribution functions and the de Hoffmann Teller velocities near the FTE seem to suggest that the FTE is generated by multiple X lines. More specifically, it could be speculated that the FTE is generated when a secondary X line, passing in proximity of the subsolar/stagnation point, develops southward of the X line at the TC 1 position. In such a case, helical field lines would be generated by magnetic reconnection occurring simultaneously at the two X lines and a flux rope would be produced (see the picture in Figure 8). Since the magnetosheath and magnetospheric fields near to the secondary X line would have approximately the same orientation as the fields near TC 1, the secondary X line would have roughly the same orientation as the X line near TC 1 and the axial direction of the helix would be parallel to the two X lines with the helix axial field being oriented as the guide field, as it is actually observed. In the proposed mechanism, the flux rope cross section increases and the internal magnetic field strengthens as new helical field lines are produced; eventually, the flux rope is convected northward by the magnetosheath plasma and TC 1 passes through it. The relative distance between the two X lines is inferred to be about 5 R E, assuming that the secondary X line passes approximately through the subsolar point and that it is parallel to the X line near TC 1. Therefore, the 3 R E extension of the FTE cross section is consistent with the proposed interpretation. [25] The formation of the secondary X line could be related to the large dipole tilt angle, as well. Raeder [2006], using a global MHD simulation, predicts the formation of flux ropes during periods when the IMF is southward and the dipole tilt is large. In the simulation, the flux ropes are generated periodically by multiple X line reconnection when a secondary X line forms near the stagnation point while the primary X line, located northward (southward) of the stagnation point for negative (positive) dipole tilt, remains active. When the magnetic shear angle is 165, the Raeder simulation predicts the generation of flux ropes with a cross section of 3 5 R E and an axial extension of R E Flux Rope Core Field [26] The FTE shows an intense core field that exceeds 60 nt, while the guide field is very small. Ma et al. [1994] studied the core field enhancement in the center of flux tubes originated by the single X line, multiple X line and patchy reconnection models through two dimensional (2 D) and three dimensional (3 D) MHD simulations. In their simulations, the magnetic field magnitudes on the two side of the current layer are equal and the magnetic shear angle is 150. Moreover, the simulations have been performed for each of two different initial magnetic field profiles across the current sheet, i.e., when the magnetic field inside the current sheet is equal to the guide field and when the magnetic field rotates across the current sheet maintaining 9of11

10 [28] Acknowledgments. This work was supported by the Agenzia Spaziale Italiana (contract ASI/INAF I/023/09/0 Attività Scientifica per l Analisi Dati Sole e Plasma Fase E2 ). We thank G. Pallocchia, M. B. Bavassano Cattaneo, and D. Ambrosino for useful discussions and E. Penou from CESR Toulouse University/CNRS who developed the CL software for the data display. We also thank both the referees for their valuable suggestions. [29] Masaki Fujimoto thanks Robert Fear and another reviewer for their assistance in evaluating this paper. Figure 8. The scheme of the flux rope in the Y GSM Z GSM plane. It must be stressed that Figure 8 only has illustrative purpose and that the X line extension toward the noon meridian cannot be deduced by TC 1 observations. the same magnitude (force free case). The results of the simulations shows that the flux ropes with the strongest core field are generated by multiple X line model in the force free field initial case. Ma et al. [1994] suggest that the reconnection at multiple X lines results in a larger magnetic tension with a consequent stronger compression which increases the core field in the flux tube. In addition, the thermal pressure gradient force along the flux rope axis accelerates plasma out of the two extremities of the tube allowing a further compression of the flux rope core field. Therefore, the enhanced core field observed inside the magnetic structure under study could indicate that the mechanism illustrated by Ma et al. [1994] is at work during the time of the observations, although it must be kept in mind that the simulation initial conditions are different from the plasma conditions on the two side of the magnetopause, in particular, the shear angle is 175 in the present case. 4. Summary [27] On 29 December 2004, TC 1 most probably stays close to an X line and therefore the orientation of such an X line can be evaluated. TC 1 observes also an FTE and, for the first time, it is found that the FTE axial direction is parallel to the direction of the X line. This indicates that the FTE is generated by a burst of reconnection at the X line or by reconnection occurring simultaneously at two reconnection lines (one close to the satellite and the other at southern latitudes) and it is not due to patchy reconnection as in the Russell and Elphic model. A series of other signatures, regarding the properties of the distribution functions and the de Hoffmann Teller velocity near the FTE, suggest that the FTE could be generated by two X lines active at the magnetopause, more consistent with the Lee and Fu FTE model of multiple reconnection X lines. More precisely, although for this event only single point observations are available, a generation mechanism similar to the one described by Raeder [2006] is envisaged. References Bavassano Cattaneo, M. B., et al. (2006), Kinetic signatures during a quasicontinuous lobe reconnection event: Cluster Ion Spectrometer (CIS) observations, J. Geophys. Res., 111, A09212, doi: /2006ja Berchem, J., and C. T. Russell (1984), Flux transfer events on the magnetopause: Spatial distribution and controlling factors, J. Geophys. Res., 89, , doi: /ja089ia08p Carr, C., et al. (2005), The Double Star magnetic field investigation: Instrument design, performance and highlights of the first years of observations, Ann. Geophys., 23, Cowley, S. W. H. (1982), The causes of convection in the Earth s magnetosphere: A review of developments during the IMS, Rev. Geophys., 20, Cowley, S. W. H. (1995), Theoretical perspectives of the magnetopause: A tutorial review, in Physics of the Magnetopause, Geophys. Monogr. Ser., vol. 90, edited by P. Song, B. U. Ø. Sonnerup, and M. F. Thomsen, pp , AGU, Washington, D. C. Cowley, S. W. H., and C. J. Owen (1989), A simple illustrative model of open flux tube motion over the dayside magnetopause, Planet. Space Sci., 37, Fairfield, D. H. (1971), Average and unusual locations of the Earth s magnetopause and bow shock, J. Geophys. Res., 76, Fear, R. C., S. E. Milan, A. N. Fazakerley, C. J. Owen, T. Asikainen, M. G. G.T.Taylor,E.A.Lucek,H.Rème,I.Dandouras,andP.W.Daly (2007), Motion of flux transfer events: A test of the Cooling model, Ann. Geophys., 25, Fuselier, S. A. (1995), Kinetic aspects of reconnection at the magnetopause, in Physics of the Magnetopause, Geophys. Monogr. Ser., vol. 90, edited by P. Song, B. U. Ö. Sonnerup, and M. F. Thomsen, pp , AGU, Washington, D. C. Gonzales, W. D., and F. S. Mozer (1974), A quantitative model for the potential resulting from reconnection with an arbitrary interplanetary magnetic field, J. Geophys. Res., 79, Gosling, J. T., M. F. Thomsen, S. J. Bame, R. C. Elphic, and C. T. Russel (1991), Observations of interplanetary and lobe magnetic field lines at the high latitude magnetopause, J. Geophys. Res., 96, 14,097 14,106. Hasegawa, H., B. U. Sonnerup, C. J. Owen, B. Klecker, G. Paschmann, A. Balogh, and H. Rème (2006), The structure of flux transfer events recovered from Cluster data, Ann. Geophys., 24, Hasegawa, H., et al. (2010), Evidence for a flux transfer event generated by multiple X-line reconnection at the magnetopause, Geophys. Res. Lett., 37, L16101, doi: /2010gl Kawano, H., and C. T. Russell (2005), Dual satellite observations of the motions of flux transfer events: Statistical analysis with ISEE 1 and 2, J. Geophys. Res., 110, A07217, doi: /2004ja Ku, C. H., and D. G. Sibeck (2000), Flux transfer events produced by the onset of merging at multiple X lines, J. Geophys. Res., 105, Lee, L. C., and Z. F. Fu (1985), A theory of magnetic flux transfer at the Earth s magnetopause, Geophys. Res. Lett., 12, Lockwood, M., and M. A. Hapgood (1998), On the cause of a magnetospheric flux transfer event, J. Geophys. Res., 103, 26,453 26,478. Ma, Z. W., A. Otto, and L. C. Lee (1994), Core magnetic field enhancement in single X line, multiple X line and patchy reconnection, J. Geophys. Res., 99, Paschmann, G., and B. U. Ö. Sonnerup (2008), Proper frame determination and walén test, in Multi Spacecraft Analysis Methods Revisited, edited by G. Paschmann and P. Daly, chapter 7, Int. Space Sci. Inst., Bern, Switz. Paschmann, G., W. Baumjohann, N. Sckopke, I. Papamastorakis, C. W. Carlson, B. U. Ö. Sonnerup, and H. Lühr (1986), The magnetopause for large magnetic shear: AMPTE/IRM observations, J. Geophys. Res., 91, 11,099 11,115. Paschmann, G., S. Haaland, B. U. Ö. Sonnerup, H. Hasegawa, E. Georgescu, B. Klecker, T. D. Phan, H. Réme, and A. Vaivads (2005), Characteristics of the near tail dawn magnetopause and boundary layer, Ann. Geophys., 23, Phan, T. D., G. Paschmann, B. U. Ö. Sonnerup (1996), Low latitude dayside magnetopause and boundary layer for high magnetic shear, J. Geophys. Res., 101, of 11

11 Phan, T. D., H. Hasegawa, M. Fujimoto, M. Oieroset, T. Mukai, R. P. Lin, and W. Paterson (2006), Simultaneous Geotail and Wind observations of reconnection at the subsolar and tail flank magnetopause, Geophys. Res. Lett., 33, L09104, doi: /2006gl Raeder, J. (2006), Flux transfer eevents: 1. Generation mechanism for strong southward IMF, Ann. Geophys., 24, Rème, H., et al. (2005), The HIA instrument on board the Tan Ce 1 Double Star near equatorial spacecraft and its first results, Ann. Geophys., 23, Russell, C. T., and R. C. Elphic (1978), Initial ISEE magnetometer results: Magnetopause observations, Space Sci. Rev., 22, , doi: / BF Scholer, M. (1988), Magnetic flux transfer at the magnetopause based on single X line bursty reconnection, Geophys. Res. Lett., 15, , doi: /gl015i004p Sonnerup, B. U. Ö. (1974), The reconnecting magnetosphere, in Magnetospheric Physics, edited by B. M. McCormac, p. 23, D. Reidel, Dordrecht, Neth. Sonnerup, B. U. Ö., and H. Hasegawa (2005), Orientation and motion of two dimensional structures in a space plasma, J. Geophys. Res., 110, A06208, doi: /2004ja Southwood, D. J., C. J. Farrugia, and M. A. Saunders (1988), What are flux transfer events?, Planet. Space Sci., 36, , doi: / (88) Trattner, K. J., J. S. Mulcock, S. M. Petrinec, and S. A. Fuselier (2007), Probing the boundary between antiparallel and component reconnection during southward interplanetary magnetic field conditions, J. Geophys. Res., 112, A08210, doi: /2007ja Trenchi, L., M. F. Marcucci, G. Pallocchia, G. Consolini, M. B. Bavassano Cattaneo, A. M. Di Lellis, H. Rème, L. Kistler, C. M. Carr, and J. B. Cao (2008), Occurrence of reconnection jets at the dayside magnetopause: Double Star observations, J. Geophys. Res., 113, A07S10, doi: / 2007JA Wang, Y. L., R. C. Elphic, B. Lavraud, M. G. G. T. Taylor, J. Birn, C. T. Russell, J. Raeder, H. Kawano, and X. X. Zhang (2006), Dependence of flux transfer events on solar wind conditions from 3 years of Cluster observations, J. Geophys. Res., 111, A04224, doi: /2005ja Zhang, H., K. K. Khurana, M. G. Kivelson, V. Angelopoulos, Z. Y. Pu, Q. G. Zong, J. Liu, and X. Z. Zhou (2008), Modeling a force free flux transfer event probed by multiple Time History of Events and Macroscale Interactions during Substorms (THEMIS) spacecraft, J. Geophys. Res., 113, A00C05, doi: /2008ja Zong, Q. G., T. A. Fritz, H. Spence, M. Dunlop, Z. Y. Pu, A. Korth, P. W. Daly, A. Balogh, and H. Reme (2003), Bursty energetic electrons confined in flux ropes in the cusp region, Planet. Space Sci., 51, , doi: /s (03) Zong, Q. G., et al. (2005), Plasmoid in the high latitude boundary/cusp region observed by Cluster, Geophys. Res. Lett., 32, L01101, doi: /2004gl J. B. Cao, School of Astronautics, Beijing University of Aeronautics and Astronautics, 1 Xueyuan Rd., Beijing, China. C. M. Carr, Department of Physics, Imperial College London, South Kensington Campus, London SW7 2AZ, UK. M. F. Marcucci and L. Trenchi, INAF Istituto di Fisica dello Spazio Interplanetario, Via del Fosso del Cavaliere 100, I Rome, Italy. (lorenzo.trenchi@ifsi roma.inaf.it) H. Rème, CESR, 9 av du Colonel Roche, BP 44346, F Toulouse CEDEX, France. 11 of 11