The reduction of source effect for reliable estimation of geomagnetic transfer functions

Abstract

SUMMERY We have analysed the literature suggestions regarding possible changes in vertical magnetic transfer function (VTF) over time. We have shown that for periods above 1500 s the observed changes in VTF are caused by the source effect and we proposed how to reduce this negative impact. For calculations we used 1-min recordings of geomagnetic variations registered between 2002 and 2017 in various geomagnetic observatories. In data processing we used frequency-domain Egbert's algorithm and our original algorithm based on the method of least squares in the time domain for some comparisons and tests. We have shown that for periods over 1500 s the VTFs calculated separately from summer and winter data are different. However, our analysis shows that the variability of the VTF values obtained is misleading and results from time-changing presence of magnetic field variations that do not fulfill the assumption of plane wave (there is a vertical component in the incident magnetic field). These variations are much more numerous in summer than in winter. More detailed analysis has shown also that they are usually small at night and big during the day. The vertical components of these variations constitute an error correlated with input signals (horizontal components), which alters the values of the determined VTF. Furthermore, error bars do not take this effect into account. It makes it impossible to improve the accuracy of calculations by increasing the amount of data. Analysing the estimated external parts of vertical components from the Central European observatories we noticed a great similarity of these signals even if the induction components were clearly different, which indicates that this is a regional effect. On this basis, we proposed a procedure to improve the accuracy of VTF determination by means of pre-selection of data.