Journal of Physical Studies 22(1), Article 1902 [4 pages] (2018) DOI: https://doi.org/10.30970/jps.22.1902 CONSTRUCTION OF A MODEL FOR PREDICTION OF THE MAGNETIC FIELD OF THE EARTH PERTURBATIONS DUE TO SOLAR ACTIVITY O. A. Baran{1}, M. I. Stodilka{1}, M. M. Koval'chuk{1}, M. B. Hirnyak{1}, I. P. Laushnyk{2} 1 The Ivan Franko National University of Lviv, Astronomical Observatory, 8, Kyrylo and Mephodij St., Lviv, UA--79005, Ukraine e-mail: sun@astro.franko.lviv.ua 2- Lviv branch of Dnipropetrovsk National University of Railway Transport, 12-a, Blazhkevich St., Lviv, UA--79055, Ukraine

A method for the short-term prediction of magnetic perturbations of the Earth as a response to changes in solar activity during an 11-year cycle has been developed. The basis of the mathematical model is the equation of linear multiple regression of the random process.

We used the time series of data obtained from the database of NOAA SWPC in the current 24th cycle of solar activity. We have studied fluxes of solar radiation, the flows of the corpuscles and the indices of geomagnetic disturbances in the Earth's magnetosphere caused by the radiation. The flows of solar wind (proton fluence with energy $ E_ {p}> 1$ MeV and electron fluence with energy $ E_ {e}> 0.8$ MeV) enhanced by X-ray streams from solar flares turned out to be the most geoeffective predictors.

Identification procedures for the prognostic model have been carried out. The input data were passed through the appropriate low-frequency Lancos filters that suppressed high-frequency noise. We found that the use of the filters improves the quality of the prediction. We established the criteria by which the quantitative characteristics of the quality and effectiveness of the results were calculated. The correlation coefficient between the observed and predicted planetary geomagnetic indices is greater than 0.9 (the use of the filters for the prediction with one day lead time increases the correlation coefficient by 10-12\%). The efficiency of the prediction with 1-3 days lead time is greater than 0.8. This indicates the validity of the developed prediction method.

PACS number(s): 94.05.S-, 95.75.Wx

pdf