Long-term evolution and prediction of geomagnetic activity

Geomagnetic activity is the disturbance of the ground-level magnetic field caused by phenomena originating from the Sun. One form of this activity are magnetic storms, which are observed most clearly at low latitudes. Many storms are preceded by sudden changes in the magnetic field (SSCs), which can also occur without a storm. Geomagnetic activity is primarily caused by coronal mass ejections (CMEs) and high-speed solar wind streams (HSSs).

This thesis studied the long-term evolution of various forms of geomagnetic activity and their relation to varying solar activity. In addition, new, longer-term prediction models for geomagnetic activity were developed, increasing the temporal span compared to prior forecast models.

Paper I investigated the yearly occurrence of magnetic storms during the space age since 1957. The study found that especially the more intense CME-storms are well-correlated with the number of sunspots, which are dark regions of strong, concentrated magnetic fields on the Sun. Less intense CME-storms are on the other hand caused also by weaker magnetic structures on the Sun. A long-term change was found in the occurrence of HSS-storms. In earlier, more active solar cycles, the peak occurrence of HSS-storms during the cycle was located close to the end of the cycle, near the sunspot minimum. In the recent, less active cycles the HSS-storm peak has shifted backward, closer to the sunspot maximum. This long-term shift can be explained by changes in the large-scale solar magnetic fields.

Paper II studied storm sudden commencements (SSCs) and their occurrence since 1868. SSCs are sudden increases of the ground magnetic field most typically seen at the beginning of intense CME-storms. The study found and corrected an inhomogeneity problem in the observation series of SSCs, which affected the number of identified SSCs in 2006-2017. SSCs are highly relevant also from a technological perspective, since the fast changes of the magnetic field associated with them induce electric currents, which can damage, for example, electric grids as well as oil and gas pipes.

Papers III and IV developed new types of longer-term prediction models for geomagnetic activity capable of predicting activity over the whole 11-year solar cycle. Earlier forecast models for geomagnetic activity have largely focused on shorter time scales of only hours to weeks. The results of this thesis can potentially be used to also improve long-term climate predictions.