ИСТИНА

The real and imaginary parts (or amplitude and phase) of physically realizable transfer functions are never independent of each other, and often turn out to be explicitly related by the integral "dispersion relation" (DR), sometimes also referred to as the "Kramers-Kronig relation", "Bode relation" and "Hilbert transform". Application of the DR may be especially effective in the magnetotelluric (MT) method, since it often allows to increase the data quality, correct the biased estimations of the impedance tensor and other MT functions, detect the inconsistent data (e.g., those heavily distorted by the source effect), etc. Unfortunately, the DR is known to be invalid in some geoelectrical situations (namely, over curved 3D conductors, strongly anisotropic bodies, and for the off-shore MT measurements), which dramatically reduces their practical value, unless the following issues are appropriately addressed: Is it possible to distinguish between the DR violations caused by noisy data and those caused by inherent properties of a measured response? Could the DR still be applied to the data obtained at the MT site where the DR were found to be invalid? In the presentation we first describe some illustrative examples of the DR application to the field MT data, and then try to get the positive answers on both of the principal questions raised above