Results of Geomagnetic Routine Observations and Earthquakes -Locality of Time Changes of Short Period Variations-
The following data and investigations are added to the previous paper (II).
(1) At Kakioka the monthly mean values (ΔZ/ΔH)'s are calculated for the year 1958. The computation of (ΔZ/ΔH)'s at Kanoya is extended to the years 1958, 1959 and 1962. The new data of (ΔZ/ΔD)'s at Memambetsu are calculated for the period 1958 to July 1962, because of the very similar shape of ΔZ with ΔD instead of ΔH. Further, (ΔZ/ΔH)'s at Simosato (φ=33°35'N, λ=135°56'E) are obtained for the years 1958 and 1959.
(2) The results given in the previous paper (II) are generally confirmed with the present data ;(ΔZ/ΔH（D))'s at any station are changed independently of those at the other stations during the same period, and their minimum, which may be responsible for reduced ΔZ's, are related to both the occurrence frequency and magnitude M of the major earthquakes near the station. Some examples showing the limit of confidence of (ΔZ/ΔH(D))'s are given in the respective tables.
(3) The quantities Ai and Bi in the experimental formula, ( ΔZ/ΔH(D))i= Ai-Bi, M, w here the suffix i expresses the station, and M the magnitude of earthquakes are given in the table 11, being determined for the whole period 1958 to July 1962.
The formula holds for the greatest earthquake with M=8 or. more. lt is interesting that the ratio Bi/ Ai is almost constant, because the ratio ΔZ/ΔH in Japan takes the maximum value in the central part of Japan and closely relates to the electric conductivity anomaly linked to the deeper part of the earth. (2) It is so desirable that the constant ratio of Bi/ Ai obtained here may be confirmed in future in several different regions of Japan and her vicinity.
(4) The maximum values of (ΔZ/ΔH(D))i keep nearly a constant level at each station, for example, 0.660 at Kakioka and 0.520 at Kanoya, even during the period in which there occurs the earthquake with M < 5.5 and further of no earthquake. Therefore, the formula (ΔZ/ΔH(D))i=Ai-Bi. M can be written in more practical expression, (ΔZ/ΔH(D))i=Ai-Bi(M-Mi'), Ai=Ai'+BiMi'. Mi' is the threshold value of M and the effect of an earthquake with M=≦Mi' upon (ΔZ/ΔH(D))i is hardly expected.
For example, taking A'κa=O.660 and A'κｙ=O.520, then M'κa=5.5 and M'κｙ=5.2 are derived. These values of Mi' are almost equal to M=5.5 which is the lowest limit of M adopted experimentally throughout the investigation.
(5) The frequencies of the minimum values of (ΔZ/ΔH(D))i in respect to the major earthquakes occurred near the i-station are given in the table 12. At Memambetsu the most of earthquakes occur in the month of minimum of (ΔZ/ΔD)ｍi', while the minimum of (ΔZ/ΔH)i precede one or two months, mainly one month, over the earthquake-month at Kakioka and Kanoya. The reason of the difference is only under a speculation at present.
(6) The Simosato observatory is situated in the region of the largest value of ΔZ/ΔH in Japan. The monthly value ΔZ/ΔH undergoes some comparable order or larger changes compared with the other stations, but there occur no corresponding earthquakes during the period concerned. It is desirable to investigate the matter for the longer period and larger earthquakes in shallow as well as deep origin.
(7) Concerning the dependence of the minimum of (ΔZ/ΔH(D))i upon the epicentral distance Δ and depth h of earthquakes, no remarkable results are obtained at present to be entrusted in future works.