Introduction

Geomagnetic storms (e.g., Gonzalez et al., 1994) are a major contributor to space weather hazards, such as disruptions of power systems and communications (e.g., MacAlester & Murtagh, 2014). During storms, solar wind energy is input from the dayside to the magnetotail and back toward the magnetosphere, which energizes the ring current that depletes the low-latitude geomagnetic field as indicated by the Dst or SYM-H indices (e.g., review by Buzulukova et al., 2018). Although the ring current is crucial to storm dynamics, there are ongoing debates about whether the ring current is powered by global circulation driven by dayside magnetopause reconnection (e.g., Dungey, 1961;Gonzalez et al., 1989) or by bursty bulk flows (BBFs) resulting from nightside magnetotail reconnection (Angelopoulos et al., 1992(Angelopoulos et al., , 1994)).

Near-Earth magnetotail reconnection (NERX) typically occurs at X ∼ -20 to -30 R E (e.g., Angelopoulos et al., 2008). Simulations (e.g., Cramer et al., 2017;Sorathia et al., 2021;Yang et al., 2016) have demonstrated that when the resulting BBFs reach geosynchronous orbit, they are an important contributor to the ring current, through associated inductive electric field, potential electric field, and impulsive transport (e.g., Keika et al., 2013). However, because of reconnection's typical distance far downtail, only a small fraction of the BBFs can reach geosynchronous orbit during either non-storm time (Dubyagin et al., 2011;Sergeev et al., 2012) or storm time (Runov et al., 2021). Closer to Earth, on the other hand, the strong dipole fields tend to suppress magnetotail reconnection (Pellat et al., 1991). Thus, although NERX is common during storms, its direct

RESEARCH LETTER

10.1029/2024GL112730 Special Collection: Space Weather Events of 2024 May 9-15 Key Points:

• During the recovery phase of the May 2024 major storm, near-Earth reconnection was likely close enough to contribute to the ring current

• Dipolarizing flux bundles penetrated to ∼6.6R E at midnight with current wedges, auroral brightening, and ionospheric currents identified

• The driver was likely a large dynamic pressure that strongly compressed the magnetosphere, causing an extremely thin current sheet at ∼6.6R E

Supporting Information:

Supporting Information may be found in the online version of this article. contribution to powering the ring current has been considered inefficient. However, recent observations by Angelopoulos et al. (2020) and Runov et al. (2022) have identified that very near-Earth reconnection (VNERX) can occur at X ∼ -8 to -13 R E during the main phase, efficiently powering the ring current. Its occurrence is attributed to the intense solar wind dynamic pressure and southward interplanetary magnetic field (IMF), enabling a thin current sheet geometry very near Earth (e.g., Sciola et al., 2023). A recent statistical study by Beyene and Angelopoulos (2024) estimated the occurrence rate of VNERX to be 1.3 events per 1,000 hr of storm main phase observations, and negligible during recovery phase.

Even though during storm recovery phase NERX may not occur earthward of X ∼ -13 R E (for it to satisfy the operational criteria of a VNERX), it may still occur closer to Earth than during non-storm times, such that it is more geoeffective, and contribute to ring current energization. On 10 May 2024, a coronal mass ejection (CME) triggered a major geomagnetic storm with the Dst and SYM-H indices exceeding -400 and -500 nT, respectively. This storm provides a good opportunity to examine whether such NERX may supply power to the ring current, which could slow down its decay during the recovery phase. Even though there was no monitor of tail flows at 13-20 R E , we present evidence that NERX could be close enough for dipolarizing flux bundles (DFBs) to penetrate geosynchronous orbit, indicating that its contribution to the ring current during the recovery phase could be more efficient than previously thought. We introduce our data set in Section 2, demonstrate our results in Section 3, and summarize them in Section 4.

Data

We utilize the Acceleration, Reconnection, Turbulence and Electrodynamics of the Moon's Interaction with the Sun (ARTEMIS) mission, which was part of the Time History of Events and Macroscale Interactions during Substorms (THEMIS) before 2010 (Angelopoulos, 2008), to monitor solar wind conditions at lunar orbit. Korean Multi-Purpose Satellite (KOMPSAT) magnetometer observations (Kim, 1999;Magnes et al., 2020) and Geostationary Operational Environmental Satellite-R (GOES-R) observations (Goodman et al., 2012) are also used to examine responses in the magnetosphere at geosynchronous orbit. We employ observations from the THEMIS ground magnetometer network (Russell et al., 2009), ground magnetometer networks obtained via the SuperMAG database (Gjerloev, 2012), and international real-time magnetic observatory network database (INTER-MAGNET; Kerridge, 2001) as well as white-light images from the THEMIS All-Sky Imager network (ASI; Donovan et al., 2006). We also use vertical total electron content (TEC) maps in North America from GPS (Hofmann-Wellenhof et al., 1998).

Results

Figure 1 provides an overview of the geomagnetic storm event on 11 May 2024, when it transitioned from the main phase to the recovery phase, as seen from the SYM-H index (Figure 1e). Within the CME sheath region, the upstream density varied very significantly from ∼10 to 80 cm -3 (Figure 1b) and the ion bulk velocity was almost 800 km/s (Figure 1c), leading to very intense dynamic pressure variations (from several to ∼30 nPa in Figure 1d). Such dynamic pressure variations caused considerable compression and expansion of the magnetosphere globally, so variations in the SYM-H index were partially correlated with the dynamic pressure variations (through the magnetopause current).

Corresponding to the dynamic pressure increase and decrease at t1-t4, the substantial back-and-forth motion of the magnetopause caused KOMPSAT on the dayside (green in Figure 1j) to temporarily enter the magnetosheath twice (Figure 1f). GOES-18, initially on the dayside (blue in Figure 1j), also briefly entered the magnetosheath at t1 (Figure 1i). Due to global expansion, GOES-16 around pre-midnight (magenta in Figure 1j) observed a local minimum in B x with strong B z at t2 (i.e., more dipole-like) followed by an increase in B x and a decrease in B z (i.e., more magnetotail-like) until t3 due to gradual compression. Similarly, local compression at t5 and significant compression at t6 resulted in local increases in the SYM-H index. Meanwhile, GOES-18 around midnight observed increases in B x and decrease in B z , while GOES-16 around dawnside observed magnetic field strength enhancements. Notably, after t5 and t6, the AE index exceeded 3,000 nT twice (Figure 1i), prompting a further examination of this time interval.

Figure 2a shows that corresponding to each extreme enhancement in the AE index, intense auroral brightening was observed by the ASI at the Athabasca station (magnetic local time (MLT) was ∼1.4 hr at 09:00 UT). The first auroral brightening (at ∼08:50 UT) expanded poleward from magnetic latitude (MLAT) of at least ∼55°-∼75°( also see Movie S1). The ASI at the Lucky Lake station, slightly east of Athabasca by ∼6°, observed eastward expansion of the auroral brightening (see Movie S2). Several minutes later at ∼08:52 UT, the ASI at the Pinawa (PIN) station (MLT of ∼2.6 hr), although partly cloudy, observed auroral brightening (see Movie S3). The second brightening (at ∼09:40 UT) was diffuse auroral brightening without substorm-like signatures (Figure 2a), so our analysis focuses on the first one. During both brightenings, ground magnetometers at high MLAT observed very significant disturbances around midnight (Figures 2b-2d; much weaker disturbances away from midnight were not shown). Especially during the first brightening, the D-component variations show sign-reversal along MLAT, consistent with current systems typical of deep injections (Yang et al., 2012)-their extremely large magnitude (4,000-6,000 nT) suggests unusually high current intensity. Additionally, VIC (MLT of ∼0.8 hr) and PIN at MLAT around the brightening onset observed a time delay of the H-component enhancement, supporting the eastward expansion of the auroral brightening (see Figure S1 in Supporting Information S1). These localized signatures are reminiscent of particle injections that could supply power to the ring current.

At middle MLAT, ground magnetometers around midnight in both southern and northern hemispheres observed enhancements in the H component (Figures 2e-2g), indicating current wedges (e.g., Kepko et al., 2014). During the first current wedge, the D-component enhancement at HON in the northern hemisphere (PPT and IPM in the southern hemisphere) suggests that the current wedge was east of HON (west of PPT and IPM), that is, centered between HON and PPT (between ∼22.5 and 23.5 hr). The identified current wedges suggest that the particle injections were likely associated with DFBs driven by magnetotail reconnection (e.g., Liu et al., 2018).

The vertical TEC map using the 1-s Global Navigation Satellite System (GNSS) receivers in North America (see Movie S4) shows that the first auroral brightening occurred to the west of Seattle. The TEC increased further over Seattle and the enhanced TEC rapidly expanded eastward, consistent with the eastward expansion of the aurora in Movies S2 and S3. The onset MLT was before 0.7 hr, consistent with that the current wedge was centered at the pre-midnight sector, and the onset MLAT was ∼53°, consistent with Figure 2a. The onset MLAT was clearly lower than usual (e.g., Gjerloev et al., 2007), suggesting that magnetotail reconnection occurred closer to Earth than typically observed during substorms (∼-20 to -30 R E ). To further investigate this, we conduct GOES-18 observations (Figure 3) whose MLT was around the first current wedge (blue in Figure 1j at t5).

Before the first vertical dotted line, GOES-18 observed a B x enhancement up to ∼250 nT relative to the international geomagnetic reference field (IGRF; Figure 3d), indicating a magnetotail-like field geometry at geosynchronous orbit with an extremely thin current sheet (as sketched in Figure 3j). Around the first vertical dotted line, GOES-18 observed dipolarization signatures, as evidenced by an increase in B z and a decrease in B x (Figure 3d; relative to IGRF model), accompanied by proton and electron flux enhancements (Figures 3e and 3f), consistent with a DFB. A few minutes later, the dipolarization signatures and particle injections were also observed at GOES-16 (magenta in Figure 1j at t5). Considering the eastward expansion of the auroral brightening, the time lag between GOES-18 and GOES-16 can be explained as eastward expansion of the substorm-like injection. These results support that NERX likely occurred close enough for the resultant DFB in the reconnection outflow to penetrate geosynchronous orbit over a wide range of MLT to supply power to the ring current at that time (It was not declared to be a VNERX because aside from the lack of equatorial spacecraft, magnetic field mapping is not accurate enough to specify the equatorial footpoint of the ionospheric activation and hence of the inferred X-line location).

Next, we discuss what might drive this NERX activity. Figure 3a shows that there was a dynamic pressure pulse around 5 min before the maximum B x at GOES-18 (Figure 3d). The 5-min delay is consistent with the propagation time from TH-C (X ∼ 41 R E , solar wind speed ∼700-800 km/s). Thus, it was likely the dynamic pressure pulse that strongly compressed the magnetosphere leading to the extremely thin current sheet at GOES-18 (sketched in Figure 3j). As upstream dynamic pressure suddenly decreased, the thin current sheet became destabilized, potentially triggering reconnection.

Later, the substorm-like signatures reached their recovery phase while the upstream dynamic pressure was rather stable. At ∼09:30 UT, the upstream dynamic pressure started to increase significantly (Figure 3a). Around 5 min later at the second vertical dotted line, GOES-18 observed that B x started to increase and B z started to decrease suggesting another current sheet thinning (Figure 3d), while the field strength at GOES-16 started to increase (Figure 3g), due to global compression. Probably because magnetotail reconnection occurred immediately after the current sheet thinning, B z at GOES-18 increased against the global compression (Figure 3d) accompanied by particle injections (Figures 3e and 3f), the second current wedge (Figures 2e-2g or Figure 3c), intense auroral brightening (Figure 2a), and significant ionospheric current (Figures 2b-2d or Figure 3b). These results suggest that another NERX event likely occurred close enough to cause a DFB to reach geosynchronous orbit and contribute to the ring current again.

Summary

During the recovery phase of the major geomagnetic storm in May 2024, we identified that near-Earth reconnection could occur close enough for DFBs to penetrate geosynchronous orbit to contribute to the ring current: (a) Intense auroral brightening and strong ionospheric currents were observed around midnight, indicative of particle injections. (b) The current wedges identified by mid-MLAT ground magnetometers around midnight indicate that the particle injections could be associated with DFBs, suggesting magnetotail reconnection. (c) The MLAT of the auroral brightening onset was very low indicating that the magnetotail reconnection likely occurred closer to Earth than typically observed during substorms. (d) GOES-18 around midnight detected DFB signatures following an extremely thin (magnetotail-like) current sheet. The driver of this process was likely the large upstream dynamic pressure that strongly compressed the magnetosphere leading to the extremely thin current sheet at geosynchronous orbit. Such upstream conditions are common during the CME suggesting that NERX might contribute to supplying power to the ring current and slowing down its decay during the recovery phase more frequently and efficiently than previously thought. Previous studies have suggested (e.g., Henderson, 2004;Skoug et al., 2003) and demonstrated (Angelopoulos et al., 2020;Beyene & Angelopoulos, 2024) that VNERX can contribute to the ring current during the main phase, yet there were no such observations linking the ring current and magnetotail signatures during the recovery phase. Our study provides further evidence from groundbased observations and the near-Earth magnetotail that NERX likely plays a significant role during storm-time recovery as well. In this study, the very intense signatures (e.g., AE more than 3,000 nT) made it relatively straightforward to identify the possible NERX events and their upstream drivers. At other times, although there could be similar NERX events that contribute to the storm, their signatures may be difficult to detect because multiple events could overlap each other embedded within background storm perturbations. For example, if the two NERX events in this study were closer to each other, it would have been very difficult to identify them separately. In the future, a statistical study could be conducted using historical and future data from multiple distances (8-15 R E , such as THEMIS on the nightside) in conjunction with geosynchronous observations and ground observations to determine the occurrence rate of NERX contributing to the ring current during the recovery phase.

Although the observed and near-Earth magnetotail signatures were very similar to those of substorms, we avoid calling them substorms but only borrow the lessons learned from how NERX phenomena progress during substorms to apply them to storm times. Principally this is because the substorm growth phase associated with energy loading and storage is absent during storms, or at least it is unrecognizable during storms due to numerous intense activations at multiple local times and latitudes which prevent a secession of AE and subsequent gradual AE increase, the hallmark of growth phase, ahead of the AE intensification related to the injection.