Identifying Flux Ropes signatures using Deep Learning
Abstract
The magnetic field configurations associated with interplanetary coronal mass ejections (ICMEs) are the main drivers of the geomagnetic activity. The prediction of such configurations is essential to Space Weather in order to forecast any resulting geomagnetic disturbances. The in-situ manifestations of the entrained magnetic uses to be associated with the magnetic flux rope topology and it is the main hypothesis for predictions.
However, it is well reported in the literature that sometimes the in-situ imprints in the magnetic field observations reveals deviations from the expected magnetic field direction rotation associated with flux ropes. It is a fact that our information about the internal magnetic structure of ICME is limited to the 1D observations of a single spacecraft crossing the large structure, leaving a considerable amount of uncertainty. This might result from changes during interplanetary (IP) evolution (see Manchester et al., 2017, and reference therein), from spacecraft crossing far from the ICME core, or possibly from the topological complexity of the magnetic structure during CME initiation in the solar corona. In this work, we carry out a sophisticated analysis of flux rope magnetic field configurations using image recognition neural networks in order to better understand the internal magnetic structure of the ICMEs. To accomplish our goal, we combine the analytical flux rope model, extracted from the physical principles, with the more advanced machine learning techniques in order to decipher the observations and evaluate the limitations of the flux rope models.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2019
- Bibcode:
- 2019AGUFMNG31A0831N
- Keywords:
-
- 1914 Data mining;
- INFORMATICS;
- 1942 Machine learning;
- INFORMATICS;
- 7599 General or miscellaneous;
- SOLAR PHYSICS;
- ASTROPHYSICS;
- AND ASTRONOMY;
- 7999 General or miscellaneous;
- SPACE WEATHER