SVD analysis of the magnetospheric AE index time series and comparison with low-dimensional chaotic dynamics

The singular value decomposition (SVD) analysis is used at different stages in this paper in order to extract useful information concerning the underl ying dynamics of the magnetospheric AE index. As a frame of reference we us e the dynamics of the Lorenz system perturbed by external noise, white or c olored. One of the critical results is that the colored noise can be differ entiated from the white noise when we study their perturbation upon the eig envalue spectrum of the trajectory matrix, the SVD reconstructed components of the original time series and other characteristics. This result is used in order to conclude the existence of strong component of colored noise in cluded in the magnetospheric AE index time series. Moreover, the study of t he SVD reconstructed components of the original time series can confirm the low-dimensionality of a dynamical system strongly perturbed by external co lored noise. Finally, the results of this study strengthen the hypothesis o f the magnetospheric chaos.