It is quite important to understand why and how extremely high amplitude waves and turbulence are generated and developed in space, since it is directly related with the behavior of background plasma as well as high energy particles. We aim to reveal the origin and the basic features of nonlinear waves and turbulence in space plasma by using theory, modeling, numerical simulations, and spacecraft data analysis.

Nonlinear evolution of MHD turbulence

The solar wind often contains various distinct magnetic field structures or waves. It is thought that such structures and waves are formed due to high nonlinearity and/or coherence in solar wind turbulence. We study how the waves have strong nonlinearity and coherence, and how they affect the processes like acceleration and diffusion of high energy particles (cosmic rays) and plasma heating, by using a variety of theoretical approaches and numerical simulations. An example of the interactions of Alfven solitons reproduced by using numerical simulation is shown below. Because of wave-wave interactions, initial Alfvenic noises are amplified to produce some solitary waves. They propagate with multiply colliding each other.

Interactions of Alfven solitons

Spacecraft data analysis using machine learning

Machine learning is a method to extract a hidden pattern in data and make a prediction. It is applied for analyzing satellite data to investigate plasma environment around the earth. We use the machine learning not only to extract some patterns which are difficult to be identified by human eyes, but also to predict high spatiotemporal resolution data from limited observational data points. The figure below shows a nonlinear wave structure called shocklet observed upstream of the earth's foreshock (upper panel) and the schematic diagram of the neural net work to identify the shocklet (lower panel). We are also doing research for making prediction of high spatiotemporal resolution plasma parameters from lower resolved observational data.