The three-hour-range index measuring geomagnetic activity


  • J. Bartels,

  • N. H. Heck,

  • H. F. Johnston


Geomagnetic activity, the frequency and intensity of magnetic disturbance, is supposed to indicate the influence of solar corpuscular radiation on the Earth. Various schemes for measuring the fluctuations of this phenomenon are in operation, but in connection with direct ionospheric studies by radio methods it was found necessary to provide a new scheme using smaller time-units than a day or half a day. A three-hour-range index K based on the “Potsdamer erdmagnetische Kennziffer” was therefore provisionally adopted by the International Association of Terrestrial Magnetism and Electricity, at Washington, D.C., U.S.A., in September 1939. Each collaborating observatory is to assign, to each three-hour interval beginning at 0h, 3h, … etc. GMT, one of the integers 0 to 9 as range-index K; their averages Km, carried to half-units, express the new measure. Scales for K are determined by assimilation of frequency-curves so that each observatory, within a year, for instance, has approximately the same number of intervals with K=0, or with K=1, etc. The principles and the practice of scaling K are described. The magnetic variations are regarded as superpositions of K-variations (to be measured), and of non-K-variations (such as regular daily variations on quiet days, to be eliminated); this is illustrated by typical cases.

For January to June 1938, tables for K and Km are given based on eight observatories. From Km, daily indices B for 24-hour intervals are derived, also equivalent ranges A. High correlations are found between K at the different observatories and Km. The effect of the persistence-tendencies on these and other correlations is studied and provides examples of general interest in statistics. In general, the correlation-coefficient r increases with the length of the time-unit used; for instance, for the daily index B and the international character-figure C, r is 0.95 for single days, but it reaches 0.98 for four-day averages.

Incidentally, a geomagnetic solar-flare effect was found preceding the outbreak of the intense storm on January 16, 1938; the time-interval suggests 22 hours as an upper limit for the travel-time of the solar corpuscles.