Open AccessRegression‐based forecast model of induced geoelectric field
Author(s) -
Lotz S. I.,
Heyns M. J.,
Cilliers P. J.
Publication year - 2017
Publication title -
space weather
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.254
H-Index - 56
ISSN - 1542-7390
DOI - 10.1002/2016sw001518
Subject(s) - field (mathematics) , solar wind , meteorology , regression , regression analysis , geophysics , geology , mathematics , statistics , magnetic field , physics , quantum mechanics , pure mathematics
The electric field induced due to solar activity is the driver of geomagnetically induced currents (GIC) in grounded conductor networks such as power grids. We present a regression‐based (i.e., empirical) model for quantitatively predicting the upper envelope of induced E field components, using near‐Earth measurements of solar wind plasma and magnetic field parameters as input arguments to the model at three midlatitude locations. Model parameters and the set of input arguments used are determined by an iterative regression process that relies on a large data set comprising selected events from 2000 to 2015. Testing on out‐of‐sample events over two years (2013 and 2015) yields correlation of between 0.68 and 0.75 when predicting the northward horizontal component of induced E field and 0.45–0.61 for the eastward component at the three stations. The forecast lead time is extended by using predicted SW parameters (from the WSA‐Enlil models) as model input. As proof of concept we predict GIC based on measured and predicted input parameters and compare that with measured GIC at a South African substation for the two St. Patrick's day events (March 2013 and 2015).