In order to obtain the characteristic variations of wind relating to the Japan-Sea coastal heavy snowfall, PCA (principal component analysis) is applied to upper-wind data and the correlations between the time-varying amplitude coefficients of PCA eigenvectors and snowfall are examined. The first and second components of wind mainly represent large-scale variations, and the several succeeding components include the meso-scale wind field besides the synoptic-scale vorticity field and vertical wind shear. Snowfall has the strongest correlation with the third component mainly representing vorticity over the southern part of the Sea of Japan. Also the fourth, fifth and sixth components have strong correlation with the snowfall in the Hokuriku District. Clear clustering of the types of heavy snowfall in Niigata Prefecture classified by Akiyama (1981a, b) is seen in the score diagram of the first and second components but not in any other component, which means that the types of heavy snowfall in Niigata Prefecture are determined largely by the large-scale wind field.
A model built in the time domain between geomagnetic variations and geoelectric field variations is applied to observation values. The parameters of the model are obtained by minimizing AIC (Akaike Information Criterion). The model with parameters obtained estimates the geoelectric field variations induced by geomagnetic variations and the residuals are calculated by subtracting the estimated values of the induced geoelectric field variations from the observation values. The result indicates that this method is available for variations with periods between a few hours and a few days. The spectral density of the residual geoelectric fields has peaks with periods of O1 and M2 constituents. On the contrary, that of the observation fields has peaks with periods of S1, S2, S3 etc.
Observations of the sea clutter with the Doppler radar were carried out at Chosi in November, 1985. The distribution of Doppler velocities undergoes a change during the passage of a cold front, and it is mainly explained by the change of the wind direction. But some part of it is attributable to the orbital velocity of the swells. Then it is properly supposed that the Doppler velocities commonly observed depend firstly on the surface wind, and secondly on the height and period of the swells. The velocities of the white-crested waves of the sea surface were measured with a video camera about 500m apart from the radar site. It is found that both the velocities of the white-crested waves and radar's scatterers on the sea surface nearly coincide. On the crest of the white-crested waves, where the sea water is mixed together with the air, the sea surface rises steeply. Its shape seems to be effective in scattering the microwave.