Estimation of solar and geomagnetic activity contribution in solar cycle variations of median peak electron density NmF2

Traditionally, it is believed that the solar cycle variations of the peak electron density NmF2 are primarily due to solar ultraviolet flux variations. Quantification of the solar ultraviolet effect is usually performed by building the linear regression of NmF2 values on the solar activity indices (F10.7 or Rz). The error of such linear regression may be associated with different reasons, particularly, with the influence of geomagnetic activity. To estimate the geomagnetic activity impact we use the double linear regression of NmF2 median values on the values of the solar (F10.7) and geomagnetic (Ap) activity indices. Selection of the optimum periods for averaging F10.7 and Ap indices is performed by minimizing the root mean square error of the linear regression. The double linear regression allows us to estimate: (1) taking into account the impact of geomagnetic activity on linear regression error, and (2) solar and geomagnetic activity contribution in solar cycle variations of NmF2 median values. In the paper we use the NmF2 values obtained with the ionosondes located at Irkutsk (52N, 104E) (2003-2015 data set) and Kaliningrad (54N, 20E) (2008-2016 data set).