Seismic instruments: article

Аnnotation: Periodic variations of various geophysical parameters (seasonal, diurnal, etc.) often have a non-sinusoidal shape. To estimate the shape of such fluctuations, the epoch superposition method (SEA) is usually used. However, in practice, the SEA estimates may be biased if the initial series contains a noticeable trend (especially non-linear) and/or a periodic component, as well as omissions of observations. In this paper, we consider the case when the omissions of observations are distributed unevenly within the period. Numerical models show that in this case, the bias of the SEA estimates persists even under the condition of preliminary detrending of the series. The paper proposes an iterative algorithm for constructing an unbiased estimate of a periodic function in such a situation. The algorithm was tested on a time series of measurements of the apparent resistance of the upper layer of the Earth’s crust in the right bank of the Chirkeyskaya HPP. The resistance of rocks was measured by the dipole electrical sounding method for two years in time increments of 1 hour. The feeding dipole was located on the surface, the receiving one at a depth of 100 m. As a result of calculations, unbiased estimates of the shape of the daily variation of electrical resistance are obtained. The amplitude of the daily variation turned out to be 1.7±0.25 %. The asymmetry of the daily effect was found: the decrease in resistance occurs about twice as fast as its growth