





Astrophysics Colloquium
Data Driven Discovery: Dynamical Networks and Statistical Analyses of Multipoint Solar Terrestrial Physics Observations
Presented by Sandra Chapman
Centre for Fusion, Space and Astrophysics, University of Warwick, UK; and Boston University Center for Space Physics
Thursday, January 18, 2018
3:00 P.M. in 169336
Abstract
The plasma and magnetic field of earth's nearspace environment is highly dynamic, with its own space weather. Space weather and solar terrestrial physics observations are increasingly becoming a data analytics challenge. Constellations of satellites observe the solar corona, the upstream solar wind and throughout earth's magnetosphere. Space weather effects on the ground are monitored by 100+ magnetometer stations in the auroral region. Ionospheric currents can be detected by magnetometers on (for example the 60+ Iridium) polar orbiting satellites in low earth orbit. These data are multipoint in space and extended in time, so in principle are ideal for study using dynamical networks. Whilst networks are in widespread use in the data analytics of societal and commercial data, there are additional challenges in their application to physical time series. Determining whether two nodes (here, ground based magnetometer stations) are connected in a network (seeing related dynamics) requires normalization w.r.t. the detailed sensitivities and dynamical responses of specific observing stations which also have seasonal variations. The spatial sampling points are not uniformly spatially distributed and are moving w.r.t. the plasmacurrent system under observation, and the plasmacurrent system itself is nonlinear and highly dynamic. This talk will present a dynamical network study of the SuperMAG set of over a hundred ground based magnetometers which observe transient dynamics of the auroral current system. Spatiotemporal patterns of correlation between the magnetometer time series can be used to form a dynamical network, the properties of the network can then be captured by (time dependent) network parameters. This offers the possibility of characterizing detailed spatiotemporal pattern by a few parameters, so that many events can then be compared with each other and with theoretical predictions.
JPL Contact: Tony Mannucci (41699)
About the Speaker
Sandra Chapman is primarily but not exclusively a plasma physicist working on problems in astrophysics and in the laboratory. She is currently Professor of Physics and Director of the Centre for Fusion, Space and Astrophysics at the University of Warwick.
Her interest in nonlinear systems began with a PhD (also at Imperial College, 1985) on chaotic charged particles in the earth's magnetosphere. This early work was recognised with the COSPAR Zeldovich Medal (commission D) and the EGS Young Scientists' Medal. She was selected to give the 2014 Royal Astronomical Society James Dungey Lecture [watch the video]. She has pioneered the development of nonlinear and complex systems approaches to solar system and laboratory plasmas and more widely, to problems outside plasma physics including climate and neuroscience. Her work using large scale numerical simulation (High Performance Computing) and modelling has included wave particle interactions, comets, plasma acceleration and heating both in the solar system, at astrophysical shocks and in magnetically confined plasmas for fusion. She has a longstanding interest in quantitative characterization of data, in particular, plasma turbulence, and her recent work develops generic data analytics methods such as network science for application to "real world" physical systems, including earth's space weather and observations of our changing climate.
She has held visiting Professorships at the Universities of Kyoto and Uppsala and was a Senior Visiting Scientist at the MaxPlanckInstitute for the Physics of Complex Systems, Dresden, the Potsdam Insitute for Climate Impact Research and an Adjunct Professor of Mathematics and Statistics at the University of Tromso.








Site Manager:
Webmaster:
CL 083220




