FORECASTING SUNSPOT NUMBERS FOR SOLAR CYCLE 25 USING AUTOREGRESSIVE MODELS FOR BOTH HEMISPHERES OF THE SUN

The SunSpot Numbers (SSNs), a proxy of the solar activity, were considered as a time series and their statistical characteristics were studied. The dynamic processes in both solar hemispheres are not strong coupled. Hence, the progress of the solar cycles was described by auto-regression (AR) models worked out for the first time separately for the Northern and the Southern hemispheres and by summation, the total SSNs were calculated. Semi-annual data were used. The model orders were determined by best approximation exante prognosis using AR models of different order to the observed solar cycle 24. A similar procedure was applied to the new solar cycle 25. The SSN maximum in the Northern hemisphere should be achieved before the maximum in the Southern hemisphere. The solar activity in the southern hemisphere would be dominant. The maximum of the total SSNs of about 117 (with a confidence interval from 77 to 165) is predicted for 2023.