Magnetic storms constitute the most remarkable large-scale phenomena of nonlinear magnetospheric dynamics. Studying the dynamical organization of macroscopic variability in terms of geomagnetic activity index data by means of complexity measures provides a promising approach for identifying the underlying processes and associated time-scales. Here, we apply a suite of characteristics from recurrence quantification analysis (RQA) and recurrence network analysis (RNA) in order to unveil some key nonlinear features of the hourly Disturbance storm-time (Dst) index during periods with magnetic storms and such of normal variability. Our results demonstrate that recurrence-based measures can serve as excellent tracers for changes in the dynamical complexity along non-stationary records of geomagnetic activity. In particular, trapping time (characterizing the typical length of "laminar phases" in the observed dynamics) and recurrence network transitivity (associated with the number of the system's effective dynamical degrees of freedom) allow for a very good discrimination between magnetic storm and quiescence phases. In general, some RQA and RNA characteristics distinguish between storm and non-storm times equally well or even better than other previously considered nonlinear characteristics like Hurst exponent or symbolic dynamics based entropy concepts. Our results point to future potentials of recurrence characteristics for unveiling temporal changes in the dynamical complexity of the magnetosphere.
Bibliographical noteThis work has been financially supported by the joint Greek-German project “Transdisciplinary assessment of dynamical complexity in magnetosphere and climate: A unified description of the nonlinear dynamics across extreme events” funded by IKY and DAAD. Individual financial support of the authors has been granted by the LINC (Learning about Interacting Networks in Climate) project (project no. 289447) funded by the Marie Curie Initial Training Network (ITN) program (FP7-PEOPLE2011-ITN), the German Federal Ministry for Science and Education (BMBF) via the Young Investigator’s Group CoSy-CC2 (grant no. 01LN1306A) and the project
GLUES, the Stordalen Foundation (Planetary Boundary Research Network PB.net), and the International Research Training Group IRTG 1740/TRP 2014/50151-0, jointly funded by the German Research Foundation (DFG, Deutsche Forschungsgemeinschaft) and the S˜ao
Paulo Research Foundation (FAPESP, Funda¸c˜ao de Amparo `a Pesquisa do Estado de S˜ao Paulo). Numerical codes used for estimating RQA and RNA properties can be found in the software package pyunicorn70, which is available at https://github.com/pik-copan/pyunicorn. The Dst data have been obtained from the World Data Center for Geomagnetism, Kyoto (http://wdc.kugi.kyoto-u.ac.jp/index.html). We are grateful to three reviewers of an earlier version of this manuscript for their detailed comments.