Probabilistic forecasting of the disturbance storm time index: An autoregressive Gaussian process approach

We present a methodology for generating probabilistic predictions for the Disturbance Storm Time(Dst) geomagnetic activity index. We focus on the One Step Ahead (OSA) prediction task and use the OMNI hourly resolution data to build our models. Our proposed methodology is based on the technique of Gaussian Process Regression (GPR). Within this framework we develop two models; Gaussian Process Auto-Regressive (GP-AR) and Gaussian Process Auto-Regressive with eXogenous inputs (GP-ARX). We also propose a criterion to aid model selection with respect to the order of auto-regressive inputs. Finally we test the performance of the GP-AR and GP-ARX models on a set of 63 geomagnetic storms between 1998 and 2006 and illustrate sample predictions with error bars for some of these events.

[1]  Aaron J. Ridley,et al.  Computational analysis of the near-Earth magnetospheric current system during two-phase decay storms , 2001 .

[2]  A. Atiya,et al.  Learning with Kernels: Support Vector Machines, Regularization, Optimization, and Beyond , 2005, IEEE Transactions on Neural Networks.

[3]  Stephen A. Billings,et al.  Data derived NARMAX Dst model , 2011 .

[4]  P. Ballatore,et al.  On the estimates of the ring current injection and decay , 2003 .

[5]  C. Russell,et al.  An empirical relationship between interplanetary conditions and Dst , 1975 .

[6]  P. Wintoft,et al.  Operational forecasts of the geomagnetic Dst index , 2002 .

[7]  N. Sheeley,et al.  Solar wind speed and coronal flux-tube expansion , 1990 .

[8]  Babak Nadjar Araabi,et al.  Locally linear neurofuzzy modeling and prediction of geomagnetic disturbances based on solar wind conditions , 2006 .

[9]  Michael W. Liemohn,et al.  Distortions of the magnetic field by storm-time current systems in Earth's magnetosphere , 2010 .

[10]  Alexander Loskutov,et al.  Testing and forecasting the time series of the solar activity by singular spectrum analysis , 2000, nlin/0010027.

[11]  J. H. Piddington.  A Hydromagnetic Theory of Geomagnetic Storms , 1962 .

[12]  Stephen A. Billings,et al.  Wavelet based non-parametric NARX models for nonlinear input–output system identification , 2006, Int. J. Syst. Sci..

[13]  Dong-Hun Lee,et al.  Comparison of Dst Forecast Models for Intense Geomagnetic Storms , 2012 .

[14]  Christopher K. I. Williams,et al.  Gaussian Processes for Machine Learning (Adaptive Computation and Machine Learning) , 2005 .

[15]  S. Billings,et al.  The analysis of electron fluxes at geosynchronous orbit employing a NARMAX approach , 2013 .

[16]  R. McPherron,et al.  Probabilistic forecasting of the Dst index , 2013 .

[17]  Alex Smola,et al.  Kernel methods in machine learning , 2007, math/0701907.

[18]  D G Krige,et al.  A statistical approach to some mine valuation and allied problems on the Witwatersrand , 2015 .

[19]  Christopher T. Russell,et al.  Long-term evolution of magnetospheric current systems during storms , 2004 .

[20]  J. Chao,et al.  Influence of the solar wind dynamic pressure on the decay and injection of the ring current , 2003 .

[21]  Johan A. K. Suykens,et al.  Coupled Simulated Annealing , 2010, IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics).

[22]  M. F. Thomsen,et al.  Plasma sheet access to the inner magnetosphere , 2001 .

[23]  S. A. Billings,et al.  Using the NARMAX OLS-ERR algorithm to obtain the most influential coupling functions that affect the evolution of the magnetosphere , 2011 .

[24]  G. Consolini,et al.  Geomagnetic D st index forecast based on IMF data only , 2006 .

[25]  R. McPherron,et al.  An empirical phase space analysis of ring current dynamics: Solar wind control of injection and decay , 2000 .

[26]  Ramkumar Bala,et al.  Real‐time prediction of magnetospheric activity using the Boyle Index , 2009 .

[27]  Robert L. McPherron,et al.  Steady magnetospheric convection and stream interfaces: Relationship over a solar cycle , 2011 .

[28]  S. A. Billings,et al.  Multi‐input data derived Dst model , 2007 .

[29]  Daniel Coca,et al.  Data derived continuous time model for the Dst dynamics , 2006 .

[30]  R. McPherron,et al.  Solar wind triggering of substorm expansion onset , 1986 .

[31]  Babak Nadjar Araabi,et al.  Multi-step prediction of Dst index using singular spectrum analysis and locally linear neurofuzzy modeling , 2006 .

[32]  Patricia H. Reiff,et al.  Empirical polar cap potentials , 1997 .

[33]  Sheng Chen,et al.  Identification of MIMO non-linear systems using a forward-regression orthogonal estimator , 1989 .

[34]  A. Berlinet,et al.  Reproducing kernel Hilbert spaces in probability and statistics , 2004 .

[35]  Jesper Gjerloev,et al.  Substorm probabilities are best predicted from solar wind speed , 2016 .

[36]  A. Yu. Loskutov,et al.  A study of the regularities in solar magnetic activity by singular spectral analysis , 2001 .

[37]  S. A. Billings,et al.  Terrestrial magnetosphere as a nonlinear resonator , 2001 .

[38]  B. Schölkopf,et al.  Kernel Methods in Machine Learning 1 , 2008 .

[39]  M. I. Pudovkin,et al.  On the Perreault-Akasofu energy function ∊. , 1986 .

[40]  J. Bartels,et al.  International data on magnetic disturbances, second quarter, 1949 , 1949 .

[41]  V. Munsami,et al.  Determination of the effects of substorms on the storm‐time ring current using neural networks , 2000 .

[42]  S. Billings,et al.  Prediction of the Dst index using multiresolution wavelet models , 2004 .

[43]  Edmund Spencer,et al.  Influence of solar wind-magnetosphere coupling functions on the Dst index , 2011 .

[44]  E. Parker,et al.  Hydromagnetic theory of geomagnetic storms , 1959 .

[45]  Duane E. Waliser,et al.  An observationally based evaluation of cloud ice water in CMIP3 and CMIP5 GCMs and contemporary reanalyses using contemporary satellite data , 2012 .

[46]  Geoffrey E. Hinton,et al.  Bayesian Learning for Neural Networks , 1995 .

[47]  J. Bartels,et al.  International data on magnetic disturbances, second quarter, 1954 , 1954 .

[48]  David G. Sibeck,et al.  Kp forecast models , 2005 .

[49]  Masahisa Sugiura,et al.  Auroral electrojet activity index AE and its universal time variations. , 1966 .