Abstract
A description is given of a technique for investigating the characteristics of atmospheric noise, and of the type of information obtained at very low frequencies. The results quoted are typical of those which have been obtained in southern England during a long period of recording, but a discussion of the statistics of all the data is not included.Atmospheric noise in a bandwidth of 300 c/s at frequencies in the range 10–35 kc/s is intermediate in character between fluctuation noise and discrete impulses. For example, the variation of the average voltage with bandwidth is similar to that for fluctuation noise, but the noise also contains peaks of high amplitude. The r.m.s. value of the envelope is of the order of five times the average value (compared with 1.1 for fluctuation noise), which illustrates the impulsive nature of the noise and shows that neither parameter by itself can provide a satisfactory description of the noise. The long-term trends, however, have been expressed in terms of the average field strength.A more detailed description of the noise is given in terms of the amplitude distribution of the peaks in the envelope and the amplitude probability distribution of the envelope itself. Either of these distributions can be expressed empirically in terms of two parameters, and can be interrelated at the higher voltage levels where the impulses have consistent shape determined by the characteristics of the receiver. At one location the amplitude of the noise may vary between wide limits with time, frequency and bandwidth, but the parameters which can be used to describe the noise structure are comparatively invariant.To supplement the amplitude data, information is required on the time sequence of the voltage changes, and it has been found that the larger impulses are randomly distributed in time. There is a tendency for the smaller impulses to occur in groups, in accordance with the knowledge that a single atmospheric often consists of a number of successive discharges, but it is doubtful whether the slight departure from randomness would have a significant effect on the severity of the interference with the usual types of radio service.