複数の観測所の資料を用い地磁気測量成果を平均静穏日夜間値に化成する方法について

抄録

The Geographical Survey Institute has established the net work of magnetic stations in Japan and the data obtained so far enable us not only to constract detailed magnetic charts for clarifying the distribution of local magnetic anomalies but also to discuss locally anoma lous geomagnetic secular variations which are becoming noticeable in connection with other geophysical phenomena, especially the occurrence of earthquake. Concerning the accuracy of epoch reduction of magnetic survey, traditional methods using the record of only one reference magnetic observatory are becoming insufficient for much precise discussion of removing the effect of transient magnetic fields from the observed values at field stations, because the old reduction methods usually neglect the geographical difference of any transient fields. In this paper a new method is proposed, in which the observed value at a field station is reduced to average night values on the presumed quiet day instead of the daily mean value, because the latter can not be regarded to have no theoretical foundation as the datum line of the transient field. At first, the night values on the actual quiet days satisfying several proposed conditions are read from the records at all available magnetic observato ries in Japan, namely, Kakioka, Kanozan, Shimosato, Memanbetsu and Kanoya. Next, these night values are expressed as quadratic function of time, by which the presumed average night values of quiet day M_n(t) can be computed on arbitrary day at these observa tories. Using these values, the geographical distribution of the deviation of actual value from the quiet night value; δM_n(t) = M(t)-M_n(t), which may be regarded the magnitude of transient field, can be obtained at the time of observation in field survey, so that the value of δM_n(t) at the magnetic field station may be determined by the geographical inter polation. Then the average night values of quiet days can be computed at any station, consequently we obtaine the magnetic datum line at the station as a regression line of the distribution of quiet night values. This method is superior against the old one both on the theoretical consistency and the accuracy of removing the transient field. The method was applied to the analysis of the 2nd Order Magnetic Survey in the Tohoku District where the observations were first made in 1954 and also repeated in 19601961. Remarka ble local anomalies of the secular variation are found especially in the northern half and along the coast line of Tohoku District.