Long-term variations of snow depth, air temperature and precipitation at Tohkamachi

Variations in the annual maximum snow depth, mean air temperature and precipitation during winter from December to February at Tohkamachi are shown in Figure 1. The maximum snow depth fluctuates widely, exceeding 3 m in some years and not reaching 1m in others. Particularly notable is the decrease in the maximum snow depth since 1988. Reference HirotaHirota and others (2006) pointed out that the timing of snow reduction in the Hokuriku area on the Sea of Japan side in the center of the main island coincided with a decrease in the amount of drift ice in the southern part of the Sea of Okhotsk and a decrease in the soil-frost depth in the Tokachi area, Hokkaido island. The coincidence was seen as possibly showing a connection between the continental-scale atmospheric circulation and specific regional phenomena such as snowfall, sea ice and soil-frost depth.

Fig. 1.

Variations in the annual maximum snow depth (a), mean air temperature (b) and precipitation (c) during winter from December to February at Tohkamachi. The solid line is the decadal mean (a). The dashed lines are approximated lines (b, c).

The winter air temperature at Tohkamachi has increased at a rate of 1.5^˚C(100 a)^–1, with a significantly high correlation coefficient of 0.41 (Fig. 1b). The relationship between the recent decrease in snow depth and increase in air temperature has been reported as described above. In this study, the variation in winter precipitation is also investigated (Fig. 1c). The decreasing trend in winter precipitation is not very distinct to date, with a correlation coefficient of –0.10. However, the inverse phase between air temperature and precipitation is obvious in the figure.

Relationship between air temperature and precipitation during winter

The relationships between mean air temperature and total precipitation from December to February at Tohkamachi for 90 winters are shown in Figure 2. A distinct negative correlation is apparent (correlation coefficient is –0.66). This implies that the total precipitation, including rain as well as snow, decreases with an increase in winter air temperature.

Fig. 2.

Relationship between mean air temperature and total precipitation during winter at Tohkamachi, 1918–2007.

To investigate the ratio between snow and rain, the relationship between precipitation and the water equivalent of snowfall is shown in Figure 3. The water equivalent of snowfall was obtained from the precipitation discriminated into snow and rain by visual observation (Reference Yamamoi, Endo, Kominami, Niwano and OzekiYamamoi and others, 2000). The inclination of the approximate line, 1.1 (nearly equal to 1.0), implies that the amount of rain changes little depending on the precipitation at Tohkamachi. Consequently, the year-to-year variation in winter precipitation depends on the amount of snowfall in the case of Tohkamachi. Similar data were obtained in a mountainous area in the southwest of Niigata Prefecture (Reference Matsuura, Matsuyama, Asano, Okamoto and TakeuchiMatsuura and others, 2005).

Fig. 3.

Relationship between winter precipitation and water equivalent of snowfall at Tohkamachi, 1922–87.

Regional characteristics of the correlation between winter air temperature and precipitation

The correlation coefficients between winter air temperature and precipitation are expected to differ among the regions. They were calculated using the long-term data of 56 manned observation sites in the snowy areas throughout Japan (Fig. 4). The open and solid circles indicate positive and negative correlation coefficients, respectively. The large circles indicate correlations higher than 0.30, and small circles indicate lower correlations. At approximately 68% of the observation sites in the snowy areas, little correlation was found, while positive correlations were found for 6 sites (11%) and negative correlations for 12 sites (21%). Most of the negative correlations were clustered in the Hokuriku area, which includes the southwest half of Niigata Prefecture on the Sea of Japan side in the center of the main island. Absolute values for this area were quite high, the highest value, –0.66, being for Tohkamachi.

Fig. 4.

Positions and correlation coefficients of 56 manned observation sites in the snowy areas of Japan. Open and solid circles indicate positive and negative correlation coefficients, respectively. Large circles indicate correlations whose absolute values are higher than 0.30. The boxed area is the area shown in Figure 6.

The mean winter precipitations at the 56 sites were plotted against the correlation coefficients for each site as shown in Figure 5. The sites located in the Hokuriku area are circled by dots. Higher negative correlations are seen for the sites with precipitation >600 mm, which are clustered in the Hokuriku area. As a huge amount of precipitation occurs in this so-called heavy-snowfall area during winter, the effects of the decrease in precipitation with the increase in air temperature are considered to become more significant.

Fig. 5.

Relationship between mean winter precipitation and correlation coefficient for each observation site. Data enclosed by dots are for sites in the Hokuriku area.

Relationship between prevailing wind direction and precipitation

Generally, a cold, dry winter monsoon blows from the Asian continent to Japan during winter. Heat and vapor are carried from the relatively warm sea surface to the atmospheric bottom layer over the Sea of Japan. This causes a decrease in the stability of the lower atmosphere, the occurrence of cumulonimbus, and heavy snowfall on the Sea of Japan side. It is well known that heavy snowfalls on the Sea of Japan side are brought during the intensification of typical winter synoptic patterns, which are characterized by high pressure over Siberia (Siberian high) and low pressure over the Pacific (Aleutian low). Cold-air advection from the north is another factor that leads to heavy snowfalls.

As stated above, the winter monsoon from the Asian continent is thought to affect the precipitation and air temperature in the heavy-snowfall areas in Japan. The relationship between the precipitation at Tohkamachi and the prevailing winds at 11 observation sites around Tohkamachi was therefore investigated using the data of the five winters 2003–07. The prevailing wind direction and mean wind speed at each site with daily precipitation more or less than 10.0 mm at Tohkamachi are shown in Figure 6a and b, respectively. Similar results were obtained at almost all sites north and west of Tohkamachi. That is to say, when high precipitation occurred at Tohkamachi, strong northwesterly winds prevailed; in contrast, when precipitation was low, the prevailing winds had a roughly southerly direction, and wind speed was small. The exceptions, such as Tohkamachi (‘To’) and Koide (‘Ko’), which are located on the riverside in the basins, are attributed to local topography (Reference Rikiishi and TojoRikiishi and Tojo, 2004).

Fig. 6.

Prevailing wind direction and mean wind speed at each observation site around Tohkamachi. Daily precipitation at Tohkamachi is (a) >10.0 mm, and (b) 0.0–10.0 mm. Ka: Kashiwazaki, To: Tohkamachi, Ko: Koide.

In Figure 7, winter precipitation and mean air temperature at Tohkamachi for the past decade (1998–2007) are plotted against the frequency of northerly and northwesterly winds at the representative coastal site Kashiwazaki (marked as ‘Ka’ in Fig. 6), respectively. The results show that precipitation at Tohkamachi increases, and air temperature decreases, with an increase in frequency of northerly and northwesterly winds. This implies that both low air temperature and high precipitation are caused by the northerly and northwesterly monsoons, and this can explain the high negative correlation between air temperature and precipitation.

Fig. 7.

Relationship between frequency of northerly and northwesterly winds at Kashiwazaki and (a) winter precipitation, and (b) mean air temperature at Tohkamachi from 1998 to 2007.