Correlation properties of a forecast of secular variation in the main geomagnetic field

Abstract

The statistical and correlation characteristics for a forecast of secular variation (SV) in the Earth’s main magnetic field were derived for the territory of Russia and the adjacent water areas in comparison with other regions of the terrestrial globe. The spatial structure of the SV forecast was found to correlate best with the SV of the preceding epoch. SV is forecasted with the largest error for the Russian territory and the adjoining equatorial area. The existence of small-size SV anomalies that vary over time in the spatial structure is characteristic for both regions. The presence of a large anomaly, such as the Brazilian Geomagnetic Anomaly, influences the error level in forecasting the mean and relative standard deviation. The absolute values of the forecast error in this case are comparable in magnitude with the values obtained for the other regions. The spatial structure of the SV forecast error points to the integral nature of SV anomalies and the independent change in the parameters of different-scale flows in the liquid core that determine them. When the methods of coefficient extrapolation are used for the SV forecast, its accuracy can be improved only by updating forecast coefficients with an interval of no more than two years on the basis of data from magnetic observatories and repeat stations.