Authors: Aatiya Ali (Georgia State University), Viacheslav Sadykov (Georgia State University), Alexander Kosovichev (New Jersey Institute of Technology)

Solar energetic particles (SEPs) and their propagation through the heliosphere and interactions with the Earth’s atmosphere result in unfavorable consequences to numerous aspects of life and technology. Given the rare nature of these events, it is crucial to study data from the Sun at different solar cycles to develop the ability to reliably predict Solar Proton Events (SPEs). In this work we report the completion of a catalog of > 10 MeV > 10 pfu SPEs observed by GOES satellites/detectors with records of their properties (start and end times, peak flux, fluence, etc.) spanning through Solar Cycles 22-24. We successfully compare the developed catalog to others; like those by NOAA’s Space Environment Services Center, for the successful validation of data processing. Using flux ratios as a proxy for spectral hardness during SPEs are explored along with the fluence distributions across various energy channels ranging from 1 MeV to 100 MeV, in addition to their relevance to the formation of ground level enhancement (GLE) events. The catalog of daily proton and soft X-ray fluxes’ statistical properties is also constructed for extending the SPE prediction effort presented by Sadykov et al. (2021) to solar cycles 22 and 23, providing a possibility to build a more robust and substantiated machine learning-driven effort. In particular, the employment of a support vector machine (SVM) classification algorithm with linear and non-linear kernels is discussed. We discuss our progress on the development of the forecasting attempt, including its extension to “all-clear” forecasts.