Superposed Epoch Analysis of Sawtooth Events: Investigating the Role of Heavy Ions for Different Interplanetary Drivers

Simulations with the Lyon-Fedder-Mobarry (LFM) global MHD model have suggested that a feedback mechanism involving heavy ionospheric ions, particularly O⁺, contributes to driving sawtooth events[1]. More recent work has suggested that this feedback is necessary for events driven by coronal mass ejections (CMEs), but that the fluctuating interplanetary magnetic field in streaming interaction region (SIR)-driven events may be sufficient to produce sawtooth events without heavy ion feedback[2]. However, a recent investigation [3] suggests that most of the O⁺ observed in the tail region at 15-19 R_E originates instead from the cusp, regardless of driver. This result is inconsistent with the feedback hypothesis, which requires the outflow to originate in the nightside auroral region. In this study we perform a superposed epoch analysis of in situ data from the tail region (-8 ≥ X_GSM ≥ -20) during sawtooth events. We find that sawtooth injections are associated with a drop in tail H⁺ density while O⁺ density is little affected; however, partial pressure due to both species decreases, suggesting that the O⁺ seen post-injection is cooler than the pre-injection population. We investigate the following questions: (1) whether there is any systematic difference between initial and subsequent injections, as predicted by the feedback hypothesis; (2) whether the behavior in CME-driven events differs substantially from that of SIR-driven events; and (3) whether the behavior is significantly different in the lobe (to which cusp outflow has direct access) than in the central and boundary plasma sheet (to which nightside outflow has direct access but cusp ions can reach only with the help of convection). We also report attempts to simulate heavy ion feedback with the BATS-R-US model, which fails to replicate the LFM results.

[1]O. J. Brambles et al. (2011), Science 332, 1183, doi:10.1126/science.1202869. [2]O. J. Brambles et al. (2013), JGR 118, 6026, doi:10.1002/jgra.50522. [3]E. J. Lund et al. (2018), JGR 123, 665, doi:10.1002/2017JA024378.