The most important parameter in the coupling between solar wind and geomagnetic activity is the $B_z$-component of the interplanetary magnetic field (IMF). However, recent studies have shown that IMF $B_y$ is an additional, independent driver of geomagnetic activity. We use here local geomagnetic indices from a large network of magnetic stations to study how IMF $B_y$ affects geomagnetic activity at different latitudes for all solar wind and, separately during coronal mass ejections (CMEs). We show that geomagnetic activity, for all solar wind, is 20% stronger for $B_y>0$ than for $B_y<0$ at subauroral latitudes of about $60^{\circ}$ corrected geomagnetic (CGM) latitude. During CMEs, the $B_y$-effect is larger, about 40%, at slightly lower latitudes of about $57^{\circ}$ (CGM) latitude. These results highlight the importance of the IMF $B_y$-component for space weather at different latitudes and must be taken into account in space weather modeling.

The interaction of the solar wind with the Earth's magnetic field produces geomagnetic activity, which is critically dependent on the orientation of the interplanetary magnetic field (IMF). Most solar wind coupling functions quantify this dependence on the IMF orientation with the so-called IMF clock angle in a way, which is symmetric with respect to the sign of the $B_y$ component. However, recent studies have suggested that the sign of $B_y$ is an additional, independent driver of high-latitude geomagnetic activity, leading to higher (weaker) geomagnetic activity in Northern Hemisphere (NH) winter for $B_y\! >\! 0$ ($B_y\! <\! 0$). In this paper we quantify the size of this explicit $B_y$-effect with respect to the solar wind coupling function, both for Northern and Southern high-latitude geomagnetic activity. We show that high-latitude geomagnetic activity is significantly (by about 40-50\%) suppressed for $B_y\!<\!0$ in NH winter and for $B_y\!>\!0$ in SH winter. When averaged over all months, high-latitude geomagnetic activity in NH is about 12\% weaker for $B_y\!<\!0$ than for $B_y\!>\!0$. The $B_y$-effect affects the westward electrojet strongly but hardly at all the eastward electrojet. We also show that the suppression of the westward electrojet in NH during $B_y\!<\!0$ maximizes when the Earth's dipole axis points towards the night sector, i.e., when the auroral region is maximally in darkness.

We use the principal component analysis (PCA) to study geomagnetic activity at annual resolution using a network of 26 magnetic stations in 1966-2015, and an extended network of 40 stations in 1980-2015. The first principal component (PC1) describes the long-term evolution of global geomagnetic activity, and has an excellent correlation with indices like the Kp/Ap index. The two networks give identical results for PC1. The second principal component (PC2) is highly correlated with the annual percentage of high-speed streams (HSS). The extended network has a slightly higher sensitivity to HSSs. We verify the non-trivial latitudinal distribution of the second empirical orthogonal function (EOF2). We find that the amplitude of the 22-year variation of geomagnetic activity has a closely similar latitudinal distribution as EOF2. This verifies that the 22-year variation of geomagnetic activity is related to HSSs. The most likely cause is the Russell-McPherron mechanism.

High speed solar wind streams (HSSs) are very efficient drivers of geomagnetic activity at high latitudes. In this paper we use a recently developed $\Delta{H}$ parameter of geomagnetic activity, calculated from the night-side hourly magnetic field measurements of the Sodankyl\"a observatory, as a proxy for solar wind (SW) speed at monthly time resolution in 1914-2014 (solar cycles 15-24). The seasonal variation in the relation between monthly $\Delta{H}$ and solar wind speed is taken into account by calculating separate regressions between $\Delta{H}$ and SW speed for each month. Thereby, we obtain a homogeneous series of proxy values for monthly solar wind speed for the last 100 years. We find that the strongest HSS-active months of each solar cycle occur in the declining phase, in years 1919, 1930, 1941, 1952, 1959, 1973, 1982, 1994 and 2003. Practically all these years are the same or adjacent to the years of annual maximum solar wind speeds. This implies that the most persistent coronal holes, lasting for several solar rotations and leading to the highest annual SW speeds, are also the sources of the highest monthly SW speeds. Accordingly, during the last 100 years, there were no coronal holes of short duration (of about one solar rotation) that would produce faster monthly (or solar rotation) averaged solar wind than the most long-living coronal holes in each solar cycle produce.

Solar coronal holes are sources of high-speed solar wind streams, which cause persistent geomagnetic activity especially at high latitudes. Here we estimate seasonal solar wind speeds at 1 AU for the last 100 years using high-latitude geomagnetic measurements and show that they give information on the long-term evolution of important structures of the solar large-scale magnetic field, such as persistent coronal holes. We find that the centennial evolution of solar wind speed at 1 AU is different for equinoxes and solstices, reflecting differences in the evolution of polar coronal hole extensions and isolated low-latitude coronal holes. Equinoctial solar wind speeds had their centennial maximum in 1952, during the declining phase of solar cycle 18, verifying that polar coronal holes had exceptionally persistent extensions just before the peak of the Grand Modern Maximum of solar activity. On the other hand, solstice speeds had their centennial maximum during the declining phase of solar cycle 23 due to large low-latitude coronal holes. A similar configuration of seasonal speeds as in cycle 23 was not found earlier, not even during the less active cycles of early 20th century. Therefore the exceptional occurrence of persistent, isolated low-latitude coronal holes in cycle 23 is not related to the absolute level of sunspot activity but, most likely, to the demise of the Grand Modern Maximum.

In the declining phase of the solar cycle, when the new-polarity fields of the solar poles are strengthened by the transport of same-signed magnetic flux from lower latitudes, the polar coronal holes expand and form non-axisymmetric extensions toward the solar equator. These extensions enhance the occurrence of high-speed solar wind streams (HSS) and related co-rotating interaction regions in the low-latitude heliosphere, and cause moderate, recurrent geomagnetic activity in the near-Earth space. Here, using a novel definition of geomagnetic activity at high (polar cap) latitudes and the longest record of magnetic observations at a polar cap station, we calculate the annually averaged solar wind speeds as proxies for the effective annual occurrence of HSS over the whole Grand Modern Maximum (GMM) from 1920s onwards. We find that a period of high annual speeds (frequent occurrence of HSS) occurs in the declining phase of each solar cycle 16-23. For most cycles the HSS activity clearly maximizes during one year, suggesting that typically only one strong activation leading to a coronal hole extension is responsible for the HSS maximum. We find that the most persistent HSS activity occurred in the declining phase of solar cycle 18. This suggests that cycle 19, which marks the sunspot maximum period of the GMM, was preceded by exceptionally strong polar fields during the previous sunspot minimum. This gives interesting support for the validity of solar dynamo theory during this dramatic period of solar magnetism.

In this paper, we study two sets of local geomagnetic indices from 26 stations using the principal component (PC) and the independent component (IC) analysis methods. We demonstrate that the annually averaged indices can be accurately represented as linear combinations of two first components with weights systematically depending on latitude. We show that the annual contributions of coronal mass ejections (CMEs) and high speed streams (HSSs) to geomagnetic activity are highly correlated with the first and second IC. The first and second ICs are also found to be very highly correlated with the strength of the interplanetary magnetic field (IMF) and the solar wind speed, respectively, because solar wind speed is the most important parameter driving geomagnetic activity during HSSs while IMF strength dominates during CMEs. These results help in better understanding the long-term driving of geomagnetic activity and in gaining information about the long-term evolution of solar wind parameters and the different solar wind structures.