A land and lake breeze is very important to understand local climate and the pattern of air pollution in the Lake Biwa basin. Detailed features of the land and lake breeze in the basin, however, have not been elucidated, due to the small water area, complicated topography and other wind systems. This paper intends to make clear occurrence frequency and geographycal distribubution of the land and lake breeze, and to discuss effects of the above mentioned factors on them, by analyzing the hourly wind data at s stations in this area (Fig. 1) during the period from January 1 to December 31, 1979.
The method of selecting the occurrence days of the land and lake breeze is as follows : In the first place, the day when an on-shore wind blew in the daytime, and an off-shore wind in the nighttime at Imazu and Hikone located in plains near the shore of the Lake Biwa is noticed. In the second place, frequency distribution of the start and the end time of the on-shore wind at Imazu and Hikone was counted (Fig. 2). In the third place, criterion for judging the land and lake breeze was determined (Table 1), and the day when the on-shore wind at Imazu and Hikone started and ended within the time ranges shown in Table 1 was determined as the occurrence day of the land and lake breeze on the shore of the Lake Biwa.
The results obtained are summarized as follows:
1. The occurrence days of the land and lake breeze in the Lake Biwa basin amounted to 60 days in warm season and 45 days in cold season in 1979 (Table 2). They were classified into two main types, type I and II (Table 4), on the basis of the general wind direction and the existance of the mountain and valley wind at Tsuchiyama in the eastern mountainous part of the Lake Biwa basin (Table 3).
2. The type I prevails when the general wind is westerly to north-westerly (Fig. 3), In this type, the land and lake breeze on the shore penetrates inland and is combined with the mountain and valley wind which has the same direction. Consequently, the combined wind diverges from the lake to the boundary of the Lake Biwa basin in the daytime and converges in the opposite direction in the nighttime (Figs. 4 and 6).
3. On the other hand, the type II is connected with the general wind of which direction is south-east to south (Fig. 3). In this type, the mountain and valley wind is not seen in the eastern part of the Lake Biwa basin. Moreover, the penetration of the land and lake breeze toward inland is limited, and effects of the general wind on the land and lake breeze is intensified in the eastern part of the Lake Biwa (Figs. 5 and 7).
4. A feature common to two types is that the lake breeze changes its direction clockwise on the eastern shore, but counter clockwise on the western shore of the Lake Biwa.

View full abstract

Precipitation distribution is in general fairly complex, because many factors are involved in it. To classify such complicated distribution a principal component analysis is a useful means. This paper aims to clarify the relationship between the variation of precipitation distribution and that of upper flow pattern, which is considered to govern the variation of precipitation distribution in winter (January and February) over North America, Monthly data during 1951_??_1978 were analysed for this purpose. Besides, some new findings with regard to the recent climatic fluctuations were obtained. The results are summarized as follows:
1. Four principal components were obtained by the principal component analysis of month ly precipitation for 38 stations (Fig. 3). The first and the second components mainly explain precipitation fluctuation of the eastern part of the United States and that of the western coast of North America respectively. Both the third and the fourth components explan more local fluctuations.
2. The values of correlation coefficients between component scores and 500 mb height prove that the first two components are closely related to the 500 mb height distribution (Fig. 5).
3. On the other hand, four components were obtained with regard to monthly mean height of 500 mb for 39 stations (Fig. 6). Both the first and the second components determine either meridional or tonal flow type (Fig. 7), and the third and the fourth components mainly explain meridional movement of 500 mb contour lines.
4. Months of high component score were chosen with respect to the first two components, e. e. four types of precipitation anomaly distribution (1 WA, 1 WB, 2 WA 2 WB) and composite maps of 500 mb height distribution were made for these four types. These maps indicate that four types of precipitation anomaly distribution are prescribed by the four different 503 erent 50 mb flow patterns respectively (Fig. 9).
5. The first two components of 500 mb height give four 500 mb flow types (1Wa 1Wb, 2Wa 2Wb ; Fig. 7). They almost coincide with the four different 500 mb flow patterns (Fig. 9) which are considered to prescribe the four different precipitation anomaly distribution types.
6. Four types of precipitation anomaly distribution correspond to four different 540 mb flow types (1 WA-1Wb, 1 WB-1 Wa, 2.WA-2Wa, 2 WB-2 Wb) . The dominant feature of 1 WA-1 Wb (Fig. 9-a) is the prevailing tonal flow with stronger middle latitude westerlees. The eastern part of the United States is a positive precipitation anomaly area. In the case of 1 WB1 Wa (Fig. 9-b), the meridional flow is dominant with a ridge over the Rocky Mountains and a trough over the eastern part of North America or the adjacent Atlantic Ocean. The eastern part of the United States is a negative precipitation anomaly area. The 2 WA-2Wa type (Fig. 9-c) is characterized by the dominant meridional flow. A ridge is formed over the area from Alaska to the Pacific Ocean and a trough is over the eastern part of North America. As for precipitation, the vicinity of 45-N of the western coast is a negative anomaly area, while Alaska and the northeastern part of Canada is a positive anomaly area. In the case of 2WB-2Wb (Fig. 9-d), the zonal flow is dominant but with weaker middle lattitude westerlies. Characteristics of precipitation distribution are contrary to the case of 2 WA-2 Wa.
7. Examination of time series of 500 mb height component scores (Fig. 8) elucidates the following tendency of flow pattern. The zonal flow was prevalent in the former half of the 1950's, whereas the meridional flow appeared in the former half of the 1960's and in the latter half of the 1970's. The circumpolar vortex has expanded southward since about 1963.
8. Judging from time series of precipitation component scores (Fig. 4), precipitation was intend to decrease in the eastern part of the United States through the investigated period.

View full abstract

The purpose of this study is to investigate the difference of rainfall distribution in relation to time-scale, in other words to clarify the difference between rainfall distribution for a short time-scale and that for a longer time-scale. The heavy rainfall of September 8_??_13, 1976, in the Shikoku island caused by typhoon 7617 Fran is chosen as a case study for this purpose. Hourly precipitation data were collected from 363 raingauge stations in the Shikoku island (Fig. 1).
Typhoon 7617 Fran affected western Japan from September 8 to 13, 1976 (Fig. 2), and more than 1, 000 mm of rain fell on the Shikoku island in a period of six days (Figs. 4 and 5, the recorded maximum total precipitation for six days was 2, 781 mm at Hiso, Tokushima prefecture). This typhoon lingered over the sea to the southwest of Kyushu from September 10 to 12 (Fig. 2), when no appreciable change in synoptic conditions took place (Fig. 3). Therefore, the period from 9 a. m. on September 10, to 9 a. m. on September 12 was chosen for this study.
Rainfall distribution maps for hourly, 3 hours, 6 hours, 12 hours and 24 hours were delineated. All maps except the hourly map were compiled by moving averages based on hourly precipitation. Axes are drawn in areas of heavy rainfall, with precipitation exceeding one-half of the maximum precipitation, which are illustrated in each chart. Summaries of the axes are presented on maps for each time-scale (Fig. 6).
The axes in a shorter time-scale (1 hour or 3 hours) are widely and randomly distributed. Any close relationship is not found between the distribution of the axes and the topography. This shows that the distribution of rainfall in this time-scale markedly changes with time, and that heavy rainfall was observed in wider areas. On the other hand, the axes in longer time-scales (12 hours or 24 hours) are unevenly distributed, and there exists an obvious relationship between the distribution of the axes and landforms. The heavy rainfall areas in such time-scales always appear along the windward side of the crests.
To conclude from the results described above, the regionality of rainfall distribution becomes clear as the time-scale increases. The distribution of rainfall in a short time-scale is not related to synoptic condition or topography; however, those in longer time-scales are controlled by synoptic condition.

View full abstract