摘要：This paper investigates the cause of annual variations in the geomagnetic total intensity that are often seen especially in the volcanic areas. As a hypothesis of the cause, a model was proposed, in which such a change is produced by changes in the inhomogeneous magnetization of near-surface rocks due to temporal changes of the atmospheric temperature. This hypothesis was tested by field and laboratory experiments. First, amplitude and phase difference of annual variations in the total intensity and ground temperature data were determined by time series analyses. Considering thermal diffusion from the surface into the ground, the phase difference between the total intensity and temperature was converted to a characteristic depth, and then the amplitude of annual temperature variation at the depth was estimated. Finally, the observed total intensity variations were compared with the expected change on the basis of the temperature dependence of rock’s magnetization obtained by a laboratory experiment and the local magnetic anomaly obtained by a magnetic survey at each magnetometer site. A good agreement between the observed and expected changes was obtained, which strongly suggests that the hypothesis is correct. It was also shown that a correction of annual variations by using temperature data will enable us more accurate detection of volcanomagnetic signals.
