Forecasting the Occurrence and Subsequent Properties of Solar Energetic Particle Events using a Data-intensive Neural Network Approach
Abstract
The Sun continuously affects the interplanetary (IP) environment by a host of interconnected and dynamic physical processes. Solar flares, Coronal Mass Ejections (CMEs), and Solar Energetic Particles (SEPs) are among the key drivers of space weather in the near-Earth environment and beyond. While some CMEs and flares are associated with intense SEPs, some show no or little SEP association. Furthermore, there is no clear and consistent connection between the properties of SEPs observed at 1 au and their progenitors at or near the Sun. The latter is due to the very complex environment that dominates SEP origin, acceleration, and transport in the IP space. To date, robust long-term (hours-days) forecasting of SEP properties (e.g., onset, peak intensities, heavy ion composition) does not effectively exist and the search for such a development continues. In this work, we present an ensemble neural-network approach in which intensive remote and in-situ data from different sources are being used to (i) forecast the occurrence of an SEP event and (ii) upon positive identification, predict its properties such as peak intensity, cumulative fluence, and the potential presence of an energetic storm particle (ESP) enhancement. We also present innovative methods for the estimation of the ensemble uncertainty, as well as the calibration of the probabilistic forecasting process; which are also being described elsewhere in this meeting. Our preliminary analysis shows that combining remote (e.g., solar images) and in-situ (e.g., real-time SW) data significantly improve the SEP occurrence probabilistic outcome and further provides a reasonable prediction of the SEP event properties.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2021
- Bibcode:
- 2021AGUFMNG41A..03D