“…For simplicity, we took one (average) frequency for the semidiurnal and dirnurnal peaks. At lower frequencies, in addition to these two peaks, we need another 3rd component (one-year periodicity) (Banno, Kuriyama 2012), since the |H(f)| 2 takes a constant value, without the 3rd component. Overall fitting of the Lorentz oscillator, including the 3rd constituent, to the practical data is reasonably good.…”