In certain observations of a downcoming radio wave, made for the purpose of measuring horizontal drifts of ionospheric irregularities, time shifts between fading curves obtained at three spaced receiving points are used to calculate the velocity and direction of drift of the amplitude pattern over the ground. Two simplifying assumptions about the pattern are frequently made, namely (i) the detail remains constant as it moves, and (ii) it is statistically isometric, i.e. the separation of two points for which the amplitudes have a given fixed correlation is independent of direction. The errors that arise if assumption (i) is wrong have been considered in a published paper. But a more general analysis, not relying on either assumption may be applied to the three fading curves, and the result expressed (either analytically or graphically) as the error of a calculation which makes the two assumptions. One important result is that if assumption (ii) is not valid the error in direction of drift may be quite serious. Thus, in an application to amplitude patterns produced by vertical reflection at 2.4 Mc/s (principally from the E layer), errors in individual results averaged 15°. Because the structure of the pattern varied from sample to sample, the error was not, in this case, systematic. But, in some applications, a preferred direction of the pattern structure could, if ignored, lead to wrong conclusions about preferred directions of drift.
Read full abstract