Can a Neural Network Model Identify Flux Rope Signatures?
Abstract
The classification monotonic and coherent magnetic configuration of internal magnetic configuration observed within Interplanetary Coronal Mass Ejections (ICMEs), often associated with a spacecraft crossing a large flux rope with helical magnetic field lines topology, is essential to predict the geomatic effect of these structures arriving at Earth. These days, we observe an increase of space- and ground-based capabilities with a growing amount of data available, which allow exploring new methodologies in our field. Recently, Dos Santos et al. (2020, Solar Physics) has carried out a study to implement ML techniques and expand our understanding of the Space Weather hazard's main drivers. Inspired by Nieves-Chinchilla et al. (2018b, 2019), we take advantage of ML techniques to interpret the ICME in situ magnetic field observations and understand in depth what in situ magnetic field observations should be expected when a spacecraft crosses flux ropes with different trajectories. We trained an image recognition artificial neural network with an analytical flux rope model. We use a pre-established Deep Neural Network handwriting model trained with synthetic data with high accuracy in well-behaved events and then evaluated against the observed ICMEs from WIND during 1995-2015. The results demonstrate that the approach works. We were able to identify flux rope signatures in 84% of simple real cases correctly and have a 76% success rate when extended to a broader set. The methodology is being improved by implementing synthetic fluctuations and adding more complex structures. These new features implemented in the model pave the way to develop a more robust tool for identifying flux rope configurations from in situ data.
- Publication:
-
43rd COSPAR Scientific Assembly. Held 28 January - 4 February
- Pub Date:
- January 2021
- Bibcode:
- 2021cosp...43E1763G