Current understanding: the sun continuously affects the interplanetary (IP) environment by a host of interconnected physical processes. Coronal mass ejections (cmes), solar flares, and solar energetic…

This is a $398,679.58 grant awarded by the National Aeronautics and Space Administration (NASA) to the Southwest Research Institute (SwRI) for solar energetic particle (SEP) forecasting research. The goal is to develop the capability to forecast SEP occurrence, onset time, peak intensities, and heavy ion properties using data-intensive studies and artificial intelligence approaches. The research will analyze relationships between solar photospheric features, eruptions, and flares, and associated interplanetary coronal mass ejections, shocks, and SEP measurements at 1 AU. The expected outcomes include a comprehensive catalog of relevant solar and interplanetary data, quantitative SEP forecasting assessments, performance metrics for AI models, and validation methods. This work directly supports NASA's Heliophysics Space Weather Operations-to-Research (H-SWO2R) program goals to improve forecasting of energetic particle conditions in the heliosphere. There is no set-aside designation mentioned for this grant award.

Current understanding: the sun continuously affects the interplanetary (IP) environment by a host of interconnected physical processes. Coronal mass ejections (cmes), solar flares, and solar energetic particles (seps) are key drivers of geomagnetic storms that could cause disruptions of technological systems such as orbiting satellites, global positioning systems, and ground-based power grids, among others. While some cmes and flares are associated with intense seps, some show no or little SEP association. Furthermore, there is no strong connection between the properties of seps observed at 1 AU and their progenitors at or near the sun. This is due to the very complex environment that dominates SEP origin, acceleration, and transport in the IP space. To date, robust long-term (hours) forecasting of SEP properties (e.g., onset, peak intensities, heavy ION composition) does not exist. Goal and science questions: our overarching goal is to determine the feasibility of forecasting SEP occurrence, onset time, peak intensities, and heavy ION properties using data intensive correlative studies and artificial intelligence approaches. We will achieve this goal by answering the following science questions: Q1. What are the relationships between the sun s photospheric spatial features, erupting regions and flare properties, and the properties of the associated icmes, IP shocks, IP plasma, and seps measured at 1 AU? Q2. Can this comprehensive set of parameters be combined to predict SEP properties at 1 AU? Data and methodology: we use x-RAY flare properties, photospheric solar images, energetic h-FE ION, plasma, and magnetic field data from SDO, soho, goes, ACE, wind, and stereo-a & b during solar cycles 23 and 24. Using ~6000 catalogued flares during 1998-2018, we will identify those associated with SEP events at 1 AU. For each event and when available, we will derive a matrix of parameters characterizing its associated flare properties, photospheric spatial features, upstream conditions, IP shock, CME, and SEP properties. Statistical and correlation studies will be performed to identify the dominant drivers that affect SEP onset times, peak intensities, and heavy ION properties. Using all inferred parameters, we will then utilize machine-learning algorithms to enhance the forecasting of SEP occurrence and physical properties at 1 AU. Expected forecast products, metrics, and validation methods: the study outcome will include the following products: - a comprehensive catalogue of solar flares and their associated photospheric features, cmes, IP plasma, and SEP properties at 1 AU. - a quantitative assessment of forecasting SEP properties at 1 AU over a time range between ~20 minutes to tens of hours. - metrics that measure the AI classifiers performances and their accuracies. We will also develop relevant skill scores to asses the goodness of each model predictability power. - algorithm to validate the models using out-of-time sampling measures. Training and testing data sets will be randomly initialized and will be cross-selected through a large iterative process. Relevance to NASA and h-SWO2R: our project responds directly to all goals of the heliophysics space weather operations-to-research (h-SWO2R) program to improve forecasts of the energetic proton and/or heavy ION conditions in the heliosphere due to solar eruptions and to improve data utilization techniques that could advance forecasting capabilities and which could also lead to improved scientific understanding. Results are relevant to two science goals of the 2012 solar and space physics decadal survey, and to a key strategic goal of NASA s heliophysics division, i.e., understand the sun and its interactions with the earth and the solar system, including space weather. The outcome also aligns with the interests of NASA space radiation analysis group (srag), NOAA space science prediction center (swpc), and the national space weather action plan (nswap).