On Possible Additional Sources of Solar Protons in the Events of September 4–10, 2017

The article considers the period from September 4 to September 10, 2017, inclusive, during which the last proton events of solar cycle 24 occurred. In order to detect possible additional proton sources and verify the sources already listed in various catalogs, we apply an empirical method for predicting proton events to all solar flares detected during this period. It is based on the threshold criteria of the parent flares. In addition, we apply an algorithm for automatic search of proton flares obtained by machine learning. Two variants of the automatic search algorithm are used: the first one (method 319) does not take into account the duration of the radio emission, while the second one (method 189) imposes a condition on its duration (>2 min). The empirical method shows that, except for the source flares found by the time of the first arrival of solar protons on Earth, other flares of this period do not fulfill all the criteria of “protonicity.” An additional test of the automatic method is the detection of proton flares that we selected by the protonicity criteria but that did not make it to the training sample. Method 319 considers proton flares X9.3 on September 6, 2017, M1.4 and X1.3 on September 7, 2017, and C8.3 on September 8, 2017, as proton flares. Method 189 does not consider the flares of September 7 and 8, 2017, as credible proton sources, which is consistent with expert empirical estimates of the protonicity criteria.