Technical Comments

Response to Comment on “Tail Reconnection Triggering Substorm Onset”

See allHide authors and affiliations

Science  12 Jun 2009:
Vol. 324, Issue 5933, pp. 1391
DOI: 10.1126/science.1168045


Lui challenges our conclusion that magnetic reconnection triggered the onset of a magnetospheric substorm. However, Lui incorrectly uses the auroral electrojet index instead of ground auroral and magnetic field pulsation signatures to determine substorm onset; single velocity and magnetic field components instead of full vectors and particle distributions to identify reconnection onset; and preliminary auroral electrojet–low index (AL) instead of ground magnometer, auroral, and magnetotail data to claim pre-existing activity.

We used data from NASA’s Time History of Events and Macroscale Interactions during Substorms (THEMIS) mission to analyze the development of three magnetospheric substorms and identified reconnection as the trigger mechanism of substorm onset (1). Lui (2) challenges our conclusion for one event on several grounds, which we address here. We stand by our original interpretation of a reconnection-initiated substorm.

To argue that substorm onset took place after the onset of near-Earth current disruption, Lui equates substorm onset with an 04:54:00 UT increase in the THEMIS auroral electrojet (AE) index [see table 1 in (1)] despite the 04:51:39 UT auroral intensification, the 04:52:00 UT onset of Pi2 pulsations (irregular pulsations in the 40- to 150-s range), and the 04:52:21 UT poleward expansion of the aurora. All of these events occurred before the 04:52:27 UT onset of near-Earth activity at probe P3 and are therefore inconsistent with the current disruption model of substorms (3, 4). Although often used as a proxy, the AE index is ill suited to identifying localized onsets.

Lui (2) further reports that the “conventional” signatures of reconnection, namely southward deflections of the magnetic field (Bz < 0, where Bz is the Z-directed component of the magnetic field in Geocentric Solar Magnetospheric coordinates) and the increase of the convective flows in the tailward direction, were not observed until after 04:53 UT. Reconnection models indeed predict these signatures, but only in the neutral sheet. Further from the neutral sheet, models predict cold ions convecting inward toward the neutral sheet superimposed upon ion beams streaming away from the reconnection site. Near the boundary separating reconnected and unreconnected magnetic field lines, reconnection models predict antiparallel beams of field-aligned low-energy electrons streaming toward, and high-energy electrons streaming away, from the reconnection site (57).

Because our study (1) established that probes P1 and P2 were not in the neutral sheet at 04:50:28 UT reconnection onset, reconnection must be identified on the basis of signatures expected outside the neutral sheet. First, P1 observed a negative, excursion of Bz (dBz < 0), P2 a positive Bz excursion (dBz > 0), and P3 no excursion—a combination that demands a reconnection topology and precludes the near-Earth plasma sheet thinning scenario. Second, the northward flow velocities dVz seen by P1 and P2 provide evidence for local reconnection inflow near these spacecraft, and their absence at P3 also excludes the near-Earth plasma sheet thinning interpretation. Third, as seen in figure 4C in (1), P1 observed the predicted tailward ion beam superimposed on a cold ion component. Fourth, as seen in figure 4E in (1), P1 observed the predicted bidirectional electron anisotropy.

Before 04:54 UT, P1 and P2 were near the separatrices within an extremely thin current sheet embedded deep within the plasma sheet. Figure 1, A and B, show the convective displacement of the dominant cold ion distributions. Because the ~20 nT magnetic field pointed Earthward along X, the dominant northward and Z-directed velocity shown in Fig. 1, C and F, was convective and consistent with the expected reconnection inflow. At P1 and P2, significant flux transport only started at 04:51 UT (Fig. 2, E and H). Figure 1B illustrates the hot, 600 km/s tailward-streaming beam expected tailward from reconnection. Cold plasma streaming anti-Sunward along the magnetic field dominated the plasma velocity moment at both spacecraft for several minutes before onset. It contributed nothing to the flux transport and is likely related to tailward motion of mantle/ionospheric plasma.

Fig. 1

(A) Ion distribution function at P2 near the time of reconnection onset in Vperp-Vpar space. (B) Same for P1. (C, F, I, and K) Convective ion flow velocity. (D, G, J, and L) ZGSM (Geocentric Solar Magnetospheric) component of the magnetic field. (E and H) Integrated (EVxB)y on P1 and P2, showing local flux transport. Vertical lines correspond to reconnection and current disruption onsets from (1).

Fig. 2

(A) Kyoto Provisional AL index, used in (2), and THEMIS AL index, including additional Canadian stations. (B) Mid-latitude Pi2 pulsation data from Pine Ridge, used in (1). (C) Total lobe pressure determined from low-energy plasma and magnetic field data on P1 and P2 satellites. (D) Same as in Fig. 1, E and H, but for a longer interval. (E and H) Ion partial flow velocity from the ESA instruments on P1 and P3, respectively. (F and H) Same as Fig. 1, C and I, but for a longer interval. (G and I) Magnetic field on P1 and P3. All vectors are in GSM coordinates; X component is blue, Y component is green, and Z component is red. Vertical lines correspond to tail reconnection onset for the two substorms, based on the times of tailward or equatorward flow.

Lui (2) points to Vz and Bz fluctuations at P1 before the 04:51:39 UT onset to refute the reconnection timing scenario. However, these fluctuations were unaccompanied by other reconnection signatures at P1 or P2 and provide negligible flux transport toward the equator (Fig. 1, E and H).

Figure 1H in (2) presents the dominant (Z) component of the convective flow at P1 on a timeline that differs from that for the other spacecraft. Figure 1 presented here shows all components together with the timing established by (1). The current disruption onset timing in (1) was based on the abrupt increases in Bz at P3 and P4, which were accompanied by high-frequency noise, consistent with (3, 4). Lui (2) defines current disruption on the basis of increases in Vx or slow changes in elevation angle, but neither is consistent with previous definitions (3, 4).

Lui employed the Kyoto Provisional AL index to imply a link between the 04:51:39 UT substorm onset and earlier activity. The THEMIS AL index, revised to include Eastern Canadian and Greenland stations and shown in Fig. 2A, shows that the earlier activity had already subsided by 04:40 UT. Neither AL index can be used to time growth phases or onset times because baseline determination requires subtraction of the monthly quiet days. Instead, we note that no mid-latitude Pi2 pulsations occurred during the 30 min before the onset (Fig. 2B). Neither P1 nor P2 observed any significant flux transport for 40 min before onset (Fig. 2D). Increasing lobe pressures inferred from the total plasma sheet pressure at P1 and P2 (Fig. 2C) during the 20 min before onset indicate growth phase flux increases, which came to an abrupt end at P2 at onset. At P3, the Bz component of P3 (Fig. 2I) decreased for 40 min before onset, the classic signature of growth phase plasma sheet thinning. Finally, the two substorms were evident in the THEMIS imager array. Aurora associated with the earlier activity had subsided to lower magnetic latitudes (~67°) and decreased in intensity (see movie S1) at least 10 min before the 04:51:39 UT onset, as expected during the substorm growth phase. All observations therefore indicate that the growth phase of the 04:51:39 UT substorm had started by 04:40 UT, and thus its onset was not due to lingering activity.

Finally, Lui (2) omits evidence for reconnection before the earlier substorm’s auroral intensification and expansion. P1 observed tailward flows at ~03:58 UT (Fig. 2E), ~6 min before the 04:04:51 UT dipolarization onset at P3 (Fig. 2I). The flows had a significant (>40 km/s) tailward convective component (Fig. 2F), followed by northward flows toward the neutral sheet, demonstrating reconnection inflow as in the 04:51:39 UT event. The images in movie S1 show auroral brightenings at 04:00:21 and 04:03:30 UT, both followed by expansions and both before current disruption at P3. These observations indicate that reconnection also triggered the earlier substorm.

In conclusion, we argue that Lui employs an antiquated substorm definition using the AE index to support a current disruption scenario. It focuses on dBz and Vx signatures at P1 to time reconnection onset, ignoring correlated signatures from two spacecraft that indicate earlier reconnection onset. Lui also presents a current disruption timing analysis that is inconsistent with previously reported definitions and employs a geomagnetic index inappropriate for studies of localized small substorms to suggest that earlier activity was responsible for the 04:51:39 UT onset. A more comprehensive index, THEMIS auroral images, and spacecraft observations reveal that reconnection-triggered earlier activity had subsided at least 10 min before the 04:51:39 UT onset, by which time the growth phase of the new substorm was well under way.

Supporting Online Material

Movie S1

References and Notes

  1. This work was supported by NASA contract NAS5-02099. The work of K.H.G. was financially supported by the German Ministerium für Wirtschaft und Technologie and the German Zentrum für Luft- und Raumfahrt under grant 50QP0402.

Stay Connected to Science


Navigate This Article