HR: 17:50h
AN: NG44A-07    [Abstracts]
TI: An Information-Theoretical Approach to Identify Nonlinearity in Magnetospheric Activity
AU: * Johnson, J R
EM: jrj@pppl.gov
AF: Princeton University, Princeton University Plasma Physics Laboratory PO Box 451, Princeton, NJ 08543, United States
AU: Wing, S
EM: simon.wing@jhuapl.edu
AF: Johns Hopkins University, Johns Hopkins University Applied Physics Laboratory, Laurel, MD 20723, United States
AB: We discuss a method to detect nonlinear dependencies in multivariate time series using mutual information and cumulant-based cost as discriminating statistics. The method is applied to the historical data stream of geomagnetic indices (which are constructed to characterize the magnetospheric state) spanning six solar cycles. The discriminating statistics of the historical data set are compared with the discriminating statistics of surrogate data streams that share the same linear properties as the historical data set. Both discriminating measures are significantly different from the surrogates a few years prior to solar minima, while no differences are apparent at the time of solar maximum. The result suggests that statistically the dynamics of the magnetosphere tend to be more linear at solar maximum than at solar minimum. We discuss how this behavior of magnetospheric dynamics may be related to the strength of the solar wind driver as well as its sensitivity to the changing dynamics of the solar wind over the course of the solar cycle. Because the strong nonlinear dependencies tend to peak on a timescale around 40-50 hours and are statistically significant up to one week, the analysis may also imply what physical processes are responsible for the nonlinear behavior. Finally, we discuss how information-theoretical methods may be used to improve predictive modeling, and we discuss the relative merits of mutual information and cumulant-based cost as discriminating statistics in the contex of limited or noisy datasets.
UR: http:w3.pppl.gov/~jrj/cumulant.html
DE: 2784 Solar wind/magnetosphere interactions
DE: 3238 Prediction (3245, 4263)
DE: 3270 Time series analysis (1872, 4277, 4475)
DE: 4430 Complex systems
DE: 4445 Nonlinear differential equations
SC: Nonlinear Geophysics [NG]
MN: 2007 Joint Assembly