Relationships Between the Spectra of Near-Earth Proton Enhancements, Hard X-Ray Bursts, and CME Speeds

Abstract

Some studies propose the transfer of flare-accelerated protons in an erupting flux rope until its reconnection with an open structure releases the trapped protons. Coulomb collisions in the dense flux-rope body deplete the low-energy part of the proton spectrum. On the other hand, shock acceleration progressively replenishes this part of the spectrum. These processes form a double power-law proton spectrum that is usually observed at the Earth’s orbit. We analyze the correlations between the slopes of near-Earth proton spectra below and above the break energy, on the one hand, and photon indices of the corresponding hard X-ray (HXR) bursts and speeds of associated coronal mass ejections (CMEs), on the other hand. We use catalogs of proton events in 1991 – 2006, HXR spectra obtained by Yohkoh and the Reuven Ramaty High-Energy Solar Spectroscopic Imager (RHESSI), and CME catalogs. Significant correlations have been found between the proton spectral slopes i) above the break energy and HXR spectral indices (0.86), and ii) below the break energy and CME speeds (−0.75). The results indicate a statistical predominance of flare acceleration at higher proton energies and shock acceleration at their lower energies. The highest-energy proton spectra reconstructed in ground-level events exhibit the second break with the steepest slope above it. Neither this slope nor the second-break energy correlates with any other parameter. This peculiarity requires understanding.

Appendix: Excluded Events

The catalogs of solar proton events composed by Logachev et al. (2016) and by Sladkova et al. (1998) contain information about probable solar sources of the near-Earth proton enhancements that is certain for the majority of events. Nevertheless, identification of some events is ambiguous. We checked these events using the online CDAW CME catalog (Yashiro et al., 2004: cdaw.gsfc.nasa.gov/CME_list/), the online RHESSI browser at sprg.ssl.berkeley.edu/~tohban/browser/?show=qli, the Yohkoh catalog composed by Sato et al. (2006) that is available in electronic form at dx.doi.org/10.1007/s11207-006-1831-5, and the Yohkoh resident database of SolarSoft, as well as other sources of information on solar events. The events with a questionable flare association were filtered out.

A number of events were excluded, because their parent flares occurred during the RHESSI or Yohkoh night or when they passed in the South-Atlantic Anomaly (SAA), where the observations were interrupted. An example that is not obvious is the SOL2006-12-13 GLE70 event (No. 46 in Table 3); almost the whole flare duration was observed by RHESSI, but the first HXR peak at 02:25, which was the strongest and hardest, was missed because of RHESSI night (Grechnev et al., 2013a). On the other hand, the Wind/Konus data terminated just before the main peak. An attempt to use the RHESSI spectrum for the second major peak at 02:29, which was also recorded incompletely, resulted in an outlier in the scatter plot in Figure 3. We therefore were forced to exclude this event. Another reason for the exclusion of an event was the flare location behind the limb (e.g. the SOL2001-04-18 GLE61 event, No. 14 in Table 4).

As mentioned in Section 2, we applied the condition $2.5 < \gamma _{\mathrm{HXR}} < 4.5$. This criterion was chosen rather arbitrarily and was confirmed in the course of the analysis. We assumed that a still softer photon index indicated meager acceleration processes, which were unlikely to provide a significant $> 50\mbox{ MeV}$ proton yield of our interest. On the other hand, a super-hard photon index does not seem realistic. In all cases when $\gamma _{\mathrm{HXR}}$ did not fit into this range, the cause was identified. We refined the time of the HXR peak in some events. As Tables 3 and 4 show, additional reasons were revealed to exclude other events with a questionable photon index, e.g. incomplete flare observations, probable data issues, etc. The situation with $\gamma _{\mathrm{HXR}} > 4.5$ along with a flare location very close to the limb indicated that the flare site could be partly occulted, which is also unacceptable.

Some events still were outliers on scatter plots. We examined each of them and identified the cause. In some cases, the spectra were non-monotonic and had a hump. Multiple changes of the operational mode occurred during some observations, as in the SOL2006-12-14 event (No. 47 in Table 3). For some questionable events, neither quick-look spectra nor quick-look images are present at the RHESSI browser. We noted all of these cases as data issues.

Table 3 presents the complete list of 47 RHESSI events excluded from the analysis. Table 4 presents a similar list of 17 excluded Yohkoh events.