2021-05-04, 15:00: Mr. Mayur Bakrania (UCL)
Mr. Mayur Bakrania
University College London, UK
Applying unsupervised learning and outlier detection methods to characterise magnetotail plasma sheet populations and instabilities
Abstract: Collisionless space plasma environments are characterised by distinct particle populations that typically do not mix. Although moments of their velocity distributions help in distinguishing different plasma regimes, the distribution functions themselves provide more comprehensive information about the plasma state. Unlike moments, however, distributions are not easily characterised by a small number of parameters, making their classification more difficult. To perform this classification, we distinguish between the different plasma regions by applying dimensionality reduction and clustering methods to electron and ion distributions in pitch angle and energy space. We test our algorithms by applying them to data from the Earth’s magnetotail. Traditionally, it is thought that the magnetotail is split into three regions that are defined by their plasma characteristics. However, we identify several more distinct groups of distributions, that are dependent upon significantly more complex plasma and field dynamics. We find clear distinctions between the electron and ion regions, showing us how each of these particle populations evolve separately.
A further use of neural networks is for outlier detection. With these methods, we identify anomalous distributions that are consistent with theoretical predictions of the tearing instability. The tearing instability is an explosive mechanism which leads to the formation of X-points and plasmoids in laboratory and space plasmas, subsequently causing magnetic reconnection. Due to its elusive nature, this is the first in-situ observation of the tearing instability. We are therefore able to characterise the temporal and spatial evolution of the tearing mode into magnetic reconnection, and how the surrounding plasma is energised during this process.