Comparative study between low and high energy SOHO/ERNE protons in solar cycle 23

We present a comparison between low and high energy solar energetic protons (SEPs) as observed by the SOHO/ERNE high energy detector in solar cycle (SC) 23. Observed and corrected peak proton intensities are used in the analyses. The linear correlations are calculated between the proton intensity and the soft X-ray class of solar flares (SFs) on one side and the projected speed of coronal mass ejections (CMEs), on another. The energy trends of these correlations are compared with previous reports.

💡 Research Summary

This study investigates the relationship between solar energetic proton (SEP) events observed by the high‑energy detector (HED) of the SOHO/ERNE instrument during Solar Cycle 23 (1996‑2009) and two solar eruptive phenomena: soft‑X‑ray flare class (ISXR) and coronal mass‑ejection (CME) projected speed (VCME). The authors focus on the lowest energy channel (HED1, 13.8‑16.9 MeV, mean ≈ 15 MeV) and the highest energy channel (HED10, 101‑131 MeV, mean ≈ 108 MeV). For each event they use the observed peak proton flux (Jp) and a corrected flux (Jp,corr) derived from a cross‑calibration with Wind/EPACT measurements. Because EPACT does not provide channels exactly at 15 MeV and 108 MeV, the correction is based on EPACT 25 MeV and 50 MeV data, yielding approximate linear relations: Jp,l,corr = ‑0.32 + 0.89 × Jp,l(EPACT) for low‑energy events (when EPACT low‑energy flux ≥ 8 DPFU) and Jp,h,corr = ‑0.37 + 0.81 × Jp,h(EPACT) for high‑energy events (EPACT high‑energy flux ≥ 1 DPFU). The authors acknowledge that these corrections are upper‑limit approximations and may over‑correct especially at high energies.

A total of 444 low‑energy SEP events were identified; 297 (67 %) could be associated with solar flares and 342 (77 %) with CMEs. Only 55 high‑energy events were available for analysis. Pearson correlation coefficients were computed between the logarithm of the proton peak flux and (i) the logarithm of the flare class (log10 ISXR) and (ii) CME speed, using both linear (log10 Jp vs VCME) and log‑log (log10 Jp vs log10 VCME) regressions. The results are summarized in Table 1 of the paper.

Low‑energy (≈15 MeV) results

• Observed flux: log10 Jp vs log10 ISXR r = 0.49 ± 0.05 (n = 297).
• Corrected flux: r = 0.50 ± 0.05 (n = 297).
• CME linear regression: r ≈ 0.55‑0.57; log‑log regression: r ≈ 0.52‑0.53.
  These values are consistent with the earlier GOES‑based study by Dierckxsens et al. (2015), which reported r ≈ 0.38 for 6 MeV protons and r ≈ 0.52 for 13 MeV protons.

High‑energy (≈108 MeV) results

• Observed flux: log10 Jp vs log10 ISXR r = 0.26 ± 0.12 (n = 52).
• CME linear regression: r = 0.36 ± 0.11; log‑log regression: r = 0.42 ± 0.10.
• Corrected fluxes show modest increases (r ≈ 0.38, 0.50, 0.52 respectively) but the uncertainties (9‑12 %) render the differences statistically insignificant.
  Importantly, the correlation between high‑energy protons and flare class is markedly lower than the 0.59 ± 0.07 reported by Dierckxsens et al. (2015) for ~117 MeV protons; the discrepancy is statistically significant.

The authors interpret these findings as follows: low‑energy SEPs display moderate positive correlations with both flares and CMEs, supporting the view that both phenomena contribute to the acceleration of particles in the ~10‑30 MeV range. In contrast, high‑energy SEPs show a stronger (though still modest) link to CME speed and a negligible link to flare class, suggesting that CME‑driven shock acceleration dominates at >100 MeV. The corrected fluxes generally increase correlation coefficients, but because the correction procedure is based on mismatched energy channels and may over‑estimate fluxes, the authors caution against over‑interpretation, especially for the high‑energy sample where only 52 events are available.

Limitations highlighted include (1) the small high‑energy sample size, (2) the approximate nature of the correction equations, and (3) potential saturation effects in the ERNE detector for very intense low‑energy events. The paper concludes that SOHO/ERNE data confirm previously reported trends for low‑energy protons but do not fully support the strong flare‑proton relationship reported for high‑energy events in earlier GOES‑based work. Ongoing work will extend the analysis to Solar Cycle 24 and incorporate more refined correction methods. The authors also plan to make the multi‑channel SEP catalog publicly available via a dedicated website.