A method of decomposition of abrupt transport phenomena in magnetized plasmas is demonstrated by using a data-driven science method. The set of turbulence data is produced for the Kelvin–Helmholtz turbulence in a cylindrical plasma by using a numerical nonlinear simulation, which is based on an extended Hasegawa–Wakatani model. The obtained turbulence state is accompanied by limit-cycle-like dynamics, and the abrupt particle transport phenomena is observed. We apply a singular value decomposition to the turbulence data with a clustering method, and the turbulence state is decomposed into background, zonal flows, and coherent and incoherent modes. The contribution of these decomposed modes to the particle transport is evaluated for all the combinations of the modes. It is found that the abrupt transport is mainly driven by the incoherent mode, and that it determines the spatio-temporal scale of the abrupt transport for the presented case.
This paper presents a method to analyze abrupt particle transport in magnetized plasmas using data-driven techniques. The authors decompose turbulence data into background, zonal flows, and coherent and incoherent modes, and evaluate their contributions to the observed abrupt transport. They find that the incoherent mode is the main driver of the abrupt transport.