Diurnal variations of precipitation from 1988 to 1993 are investigated using the radar echo composite chart covering the Japanese archipelago and adjacent seas at intervals of one or three hours. We focus on studying expansive areas on land and sea connected by eighteen digitized radars, and determining variations of precipitation due to the following cloud categories based on a horizontal scale and maximum precipitation intensity:large-scale precipitation clouds originating in quasi-horizontal circulation and cumulus-scale precipitation clouds originating in convection. The cumulus-scale precipitation cloud, defined as small precipitation clouds with strong intensity, predominates in afternoon peaks over four major islands of Japan excluding some points, the south Korean Peninsula, and Okinawa at about 26N. This is attributed to boundary layer heating by solar radiation during the daytime.Other precipitation clouds mainly composed of the large-scale precipitation cloud exhibit a diurnal cycle peaking in the morning, regardless of land-sea circulation.Westerly wind fields over Japan have two effects on diurnal variations.Westerly winds over northeastern Japan maximize the cumulus-scale precipitation cloud over the east; and cause an evening peak over the eastern inshore by driving clouds east. Large-scale precipitation cloud driven by westerly wind often pass through the Nansei Islands from the northwest in the afternoon between May and July, and are responsible for afternoon peak precipitation in these islands.

The effect of the passage of solar sector boundaries (SSB) on the variation of zonal mean fields below the 100mb level is investigated for the period from November 1963 to March 1978. From a key-date analysis it is found that the polar temperature in the troposphere reaches a minimum on the passage date of SSBs during winter in the period from December 1964 to February 1972. At the 500mb level of the North Pole, the filtered temperature on the passage date deviates -2.59K from the mean. Such a large deviation as -2.59K was not obtained in the random sampling of key dates. This response appears in each of 8 winters from 1965 to 1972 and each of 3 months in winter.

A radar echo composite chart (RECC) displays precipitation intensity levels in 5 Â 5-km grids for a composite region of digitized radars covering more than 2 Â 10 6 km 2 . Each continuous precipitation cluster in a RECC is considered a precipitating cloud (p-cloud).At around 400 km, the distribution of the horizontal scale (S) for p-clouds changes. At scales smaller than 400 km, the distribution decreases smoothly with scale. In the distributions of S and maximum precipitation intensity (A) for p-clouds, there is a peak at between 400 and 1,000 km and about 45 mm/h throughout the year. This peak results from large-scale p-clouds accompanying macro-b and meso-a-scale disturbances.From June to September over land, precipitation is concentrated at scales between 15 and 200 km and intensities exceeding 8 mm/h. This distribution results from cumulus-scale p-clouds, which include a variety of convective clouds initiated by boundary-layer processes. A third type of p-cloud, caused by air-sea interactions, appears with the cold northwesterly flow that follows the passage of extratropical cyclones, especially in winter. This type dominates the weaker intensity side of the cumulusscale p-clouds between 5 and 200 km in the precipitation distribution.

The aim of this study is to find a linkage of atmospheric responses to the solar sector boundary (SSB) passages (Wilcox et al., 1974;Misumi, 1981;1983). Variations of both zonal mean fields and planetary waves around the SSB passages are investigated using the transformed Eulerian mean (TEM) equations system including the refractive index. Deviations from the time mean field in 8 winters from 1964 to 1972 are mainly discussed.It is found that the Eliassen-Palm (EP) flux for zonal wavenumber 2 and 3 in the troposphere begins to decrease about 5 days prior to SSB passages. Such decrease leads to the divergence of the EP flux near the tropopause in high latitudes. This is responsible for the temperature decrease in the polar troposphere, and consequently for the decrease of the vorticity area index in the upper troposphere.The decreases of the EP flux for both wavenumber 2 originating from the troposphere and wavenumber 1 being at the 150mb level on 4 days before the SSB passages seem to create the minimum of the polar temperature in the middle stratosphere.Among above mentioned processes, only the decrease of the EP flux is not diagnosed by the TEM equations system.The decrease of the EP flux in the troposphere is statistically significant and is seen only in the same period that other atmospheric responses occur. This type response would be due to the external modulation.