The results of synoptic analysis are described concerning the two typical examples of winter-time polar lows over the Japan Sea area. That occurred in the period of 24-27th FEB. 1981.
   In this report, the following results are obtained.
   (1). Non-frontal cyclones of small dimensions are formed in polar air masses over the Japan Sea in winter. They are located on the forward sides of the positive vorticity maxima at 500 mb and accompanied with a spiral or comma-shaped cloud pattern as they mature. They are also associated with the convectively neutral layers through a substantial depth of the troposphere and situated in regions marked by well-developed baroclinity throughout the troposphere in the poleward side of the jet stream. These features show that these cyclones are so-called polar lows.
   (2). There are two-kinds of vorticity maxima at 500 mb. One is located in the central regions of cold vortex, and the other in the trough areas. The surface cyclones combined with the cold vortices are also of small dimensions below about 1000 km over the Japan Sea area and can be regarded as polar lows.
   (3). The formation of deep layers of convective neutrality is investigated. It is shown by the stability change equation that the convectively neutrae layers of the middle troposphere between 700 and 500 mb on the forward areas of the positive vorticity maxima at 500 mb are mainly formed by the horizontal convergence effect, while in the rear areas of these maxima the layers are stabilized by the horizontal divergence effect. The layers below 700 mb are destabilized by low-level heating over the oceans in winter. These two layers combine and form the convectively neutrae layers through a considerable depth of the troposphere in front of the upper positive vorticity maxima.
   (4). Instability mechanism for the formation of these cyclones are discussed by using of Petterssen's development equation and the distributions of Richardson numbers. It is suggested that the baroclinity effect, including the influence of the vorticity advection in the upper atmosphere on the surface cyclogenesis, and the non-adiabatic effect in the atmosphere of small Richardson numbers below about 10 are the main cause of the formation of the cyclones of small size below about 1000 km.
   (5). The heavy snowfalls caused by the convective cloud bands accompanied with these polar lows in the coastal plain regions along the Japan Sea are also investigated. It is shown that the distributions of daily precipitation amounts are interpretable by the distributions of these convective cloud bands.

View full abstract

   A continuous gravity observation system for the purpose of detection of minor gravity changes associated with volcanic activities through the analysis of earth tide data has been operated since late March, 1981, inside the seismology and gravity observation building (Figs. 1, 2 and 3 and Photo 1) of the Meteorological Research Institute (36°03′08" N, 140°07′48" E and 21 m a.s.l.), Tsukuba, Japan. Its sensor is No. 17 LaCoste & Romberg ET (Earth Tide) Gravity Meter of null-method type, having a higher resolution than 1 μgal and a servo detection system (Table 1 and Fig. 4).
   In this paper, the stability of the system and the response characteristics to variations of atmospheric pressure, base tilts and ambient air temperature are discussed by analysis of continuous observation materials (hourly data) through early April, 1984.
   From the observed gravity data, 4 major tidal components (O₁, K₁, M₂ and S₂) are well separated (Fig. 5). Drift of gravity data is removed from the observed data by using a digital filter proposed by Nakagawa (1961), and tidal data of the observed materials are synthesized with tidal parameters (the maximum number is 21) by means of the least square method as shown in Fig. 6. The G-factor and the phase lag of each tidal component are derived by comparison between the observed data corrected for atmospheric pressure and the theoretical tide by the method of Tamura (1982). Responses of the gravity meter against environmental variations such as atmospheric pressure are also analyzed from individual correlations.
   The drift rate exceeding 25 μgal/day after the installation of the system was observed, but the rate decreased to the designed value of less than 300 μgal/month after 6.5 months. The drift rate since the reinstallation in June 1982 shows a rather complex trend, but dropped to several microgals per day in early 1984 (Fig. 7). The phase of the seasonal component of gravity meter drift is similar to that of ambient air temperature (Table 2).
   Seasonal variations are also recognized in the data of atmospheric pressure, base tilts and ambient air temperature (Fig. 8). A high correlation between gravity and atmospheric pressure data is obtained for a range longer than diurnal period, but in cases of base tilts and ambient air temperature, there are not high correlations between them and gravity (Table 3). There is no effective change in the relative gravimetric sensitivity during the observation period, estimated by the method of Nakai (1977) (Fig. 9). A smooth curve of drift of gravity data is obtained by corrections for atmospheric pressure and ambient air temperature (Fig. 10).
   The correction factor between gravity and atmospheric pressure data is estimated to be -0.34 μgal/mb (Table 3). There is no remarkable change of this factor throughout the observation period. Almost the same value is obtained from a change of gravity data associated with a sudden change of atmospheric pressure when the typhoon passed off the Kanto area, Japan (Fig. 11).
   The G-factors and the phase lags of 4 major tidal components (O₁, K₁, M₂ and S₂) from gravity data show stable values throughout the period, except K₁ component which has large seasonal variation (Fig. 12). The G-factor of M₂ component is obtained to be 1.209 (Table 4).
---- The future omission(over 4000 character), Please see PDF file ----

View full abstract