Premium
Linear prediction filters for linear and nonlinear modeled geomagnetic activity
Author(s) -
Klimas A. J.,
Baker D. N.,
Roberts D. A.
Publication year - 1991
Publication title -
geophysical research letters
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.007
H-Index - 273
eISSN - 1944-8007
pISSN - 0094-8276
DOI - 10.1029/91gl01148
Subject(s) - asymptote , nonlinear system , linear filter , control theory (sociology) , frequency response , earth's magnetic field , amplitude , linear model , physics , mechanics , filter (signal processing) , mathematics , mathematical analysis , computer science , optics , engineering , quantum mechanics , magnetic field , statistics , control (management) , artificial intelligence , electrical engineering , computer vision
It is shown that the Faraday loop analogue model of geomagnetic activity exhibits both the directly driven and loading‐unloading magnetospheric responses to solar wind input. It is further shown that the directly driven component is a linear response to loading while the loading‐unloading response is nonlinear. Linear prediction filters which relate model input to output are discussed. By either allowing or suppressing the loading‐unloading model response, filters that relate to nonlinear or linear dynamics, respectively, have been computed. Filters that describe the directly driven response are finite ranged; they asymptote to zero with increasing lag on a time scale that is fixed by the dissipation rate of the model. Filters that describe the nonlinear total response are infinite ranged; they asymptote with increasing lag to large amplitude periodic oscillations. Some implications of these infinite ranged filters are discussed.