Introduction

Snow and ice are good indicators or the coldness of a given winter and area. Variations of snow and ice amounts and accumulation are a measure or local and global climate changes. including warming, and have been discussed from a global- change point of view (Reference MeierMeier, 1984: Reference Houghton, Jenkins and EphraumsHoughton and others, 1990; Reference Hall, williams and BayrHall and others, 1992; Reference Barry, Goodison and LeDrewBarry and others, 1993). In addition, apart from being a water resource. snow is important for the disastrous human consequences it can have.

The Hokuriku district or Honshu island, Japan, has suffered from heavy snowfalls (“Gosetsu” in.Japanese) in the past such as 36. 38, 56, 59. 60 and 6l Gosetsu winters, In a Gosetsu wintersIn a Gosetsu winter, snow reaches maximum depths of more than 4 m on the ground in Tokamachi, a city of 50000 people in Hokuriku district. In Nagaoka in the 1963 winter. the maximum snow depth on the ground reached 3.18 m. In these extreme winters. hundreds or people may be killed because or heavy snowfalls. For example, in the 1963 winter. 231 persons were killed (Reference Takahashi and NakamuraTakahashi and Nakamura, 1986), Therefore, residents or Hokuriku district. as well as of other snow-covered areas of Japan, need to have timely and accurate predictions or snowfall.

Geographic Character of Nagaoka

Nagaoka (37°25′ 97 m.a.s.l) is located in the northern part of the Hokuriku district 0Honshu island. Storms, which produce heavy snowfalls. develop Over the Sea of Japan, whichc is the water-vapor source for winter storm systems. In Hokuriku district, there is usually no wind when heavy, snowfull occur, The distance from the nearest coast to (20km kill away) to mainland. Asia is about 1000 km. the longest of any part or the Sea or Japan (Reference Hozumi and Magono.Hozumi. and Magono, 1984) Nagaoka is located close to 1000 m high mountains. in the northernmost area of the heavy-snowfal1 district. The winter air temperature is the coldest in the district, and snow as a percentage of the total precipitation in winter is the highest in the district. Thererore, Nagaoka offers one of the best locations to investigate snow amount as an indicator or global change, Figure 1 shows annual and daily changes or snows over measured on the ground at Naagaoka between the winters or 1935-36 and 1993-94, (Data from two winters were added to those from the Nagaoka City Office (1992)). Each pattern corresponds to one winterseason, The area under the cruve indicators the amount snowfall in a given winter. The maximum snow depth is also a measure of the amount of Snow as seen in the figure. The maximum snow depth occurs depth around February except in the 1945 and 1984 winters, In this papn the maximum snow depth on the ground is used as an indicator of the snow amount.

Fig. 1. Annual and daily changes of snow cover on the ground at Nagakoa, Hokuriku district (1935-36 to 1993-94 winters)Each Pattern corresponds to one winter

Data Source

Data for the maximum snow depth analyzed in this paper were taken from the paper by Reference Ikarashi, Hayakawa and KanckoIkarashi ancl others (1992), in which the maximum snow-depth data is Nagaoka from 1989 to 1992 were compiled, For the present study only the snow data set from 1905 to 1994 was analyzed due to the lack of air-temprature data from 1989 to 1904 was, Precipitation and air-temprature data for the winters from Decemcer to February 1905-94 were collected at the Japanese Meteorological Agency in Niigata Which has gathered and compiled them.

Snow and Winter Air Temperature, 1982-94

Figure 2 shows the observed maximum snow depth on the ground, precipitation in winter (December-February) and winter mean air temprature from 1982 to 1994. It shows that in the last 13 winters, precipitation including both rain and snow has not varied greatly, though the maximum snow depth decrease drasticallY. This drastic decrease seems to be a due to winter mean, air-temperature rise as shown in the figure.

Fig. 2. Temporal varitions in the maximum snow depth (cm), precipitation (mm) and mean air temprature (°C) of winters (December-February), 1982-84.

Snow and Winter Air Temperature in the Last Century

Figure 3 shows annual variation of the maximum snow depth measured on the ground from 1905 to 1994 In the past century the maximum depth of 3.18 m was observed in the winter of 1963 (38 Gosetsu year) with in a mean depth of 1.27 m and a standard deviation of 0.62 m A smoothed curve shows 10 years moving average and was constructed to shows decdal and longer and time-scale trend more clearly These 10 year averages were calculated from values in the four previous years, the present year and the five following years. Figure 4 shows annual Variation of the winter (Deccember-February) precipitation. Thc maximum value of 1516.5 mm was found in 1945. with a mean of 951.8 mm and standard Variations of 243.5 mm. A smoothed curve is a also developed for total winter precipitation. Figure 5 shows annual variation of the winter (December-February) mean air temperatures. The maximum a nd minimum mcan a ir temperatures were observed in 1949 and 1945. respectively.

Fig. 3. Temporal varitions of the maximum snow deoth (m), 1905-94 A smooth curve is a filtered value designed to show decadel and longer time-scale trends more clearly.

Fig. 4. Temporal varitions of precipitation (mm) in winter (December-February) 1905-94. A smooth curve is a filtered value designed to show decadel and longer time-scale trends more clearly.

Fig. 5. Temporal varitions of mean air temprature (°C) in winter (December-February) 1905-94. A smooth curve is a filtered value designed to show decadal and longer time-scale trends more clearly.

The a nnual winter mean air temperature in the past century as 1.6°C with a standard deviation of 1.0°C. A smooth curved is also developed for mean winter air temperature as for maximum snOw depth and precipitation. The minimum or the smoothed cruve of mean air temperature occurred in 1942 as shown Figure 5. Figure 6 shows annual variation of the deadal filtered value (10 year moving average) of the winter mean air temperature with a regression line of the filtered values obtained from 1905 to 1994. The annual variation of 10 year moving average shows a periodic pattern similar to a pattern in the Northern Hemisphere produced by Reference Houghton, Jenkins and EphraumsHoughton and others (1990). Tile equation of the regression line with a correlation ion coefficient r was expressed as:

where T is winter mean air temperature in °C and t is years AD. The increase in winter mean air temprature of 1.35°C in the past century is rather large. An airtemperature rise calculated by using only two means (average data ), of the decades 1905-14 and 1985-94, was obtained as 0.79°C.

Fig. 6. variation of mean air temprature (December-February) filtered with decades with a monotonic trend line.

Correlation bctween the maximum snow depth and the winter mean air temperature measured on the ground at Nagaoka is Shown in Figure 7. The figure shows that there is an inverse negative correlatio between the maximum snow depth and the mean air temperature. The regression equation with a correlation coefficient r was ex p resscd as :

where S is the maximum Snow depth in cm, and T winter mean air temprature in °C.

Fig. 7. Correlation between maximum snow depth (m) and mean air temprature (°C) in winter (December-February), 1905-94.

Figure 8 shows, a positive correlation between the winter precipitation and the maximum snow depth. The regression line with a correlation coefficient r was expressed as:

where P is the precipitation in mm and S is the maximum Snow dcpth in Cm. As shown in Figure 9, winter precipitation decreases as the winter mean air temperature increasrs, The regression line with a correlation coefficiellt r was obtained as:

where P is mm, and T mean air temperature in °C. Figure 10 Shows annual variation of hree decadally filtered values of maximum snow depth, winter mean air tcmperature and winter precipitation. As shown in Figure 10, both the maximum Snow depth (1.8 m) and maximum winter prceipitatioll (1200 mm) were found in 1910 with the minimum mean air temprature of 0.4°C occuring in 1942.

Fig. 8. Correlation between precipitation (mm) in winter (December-February) and maximum snow depth (m), 1905-94.

Fig. 9. Correlation between precipitation (mm) and mean air temprature (°C) in winter (December-February) 1905-94.

Fig. 10. Inverse postive correlation between maximum snow depth on the ground (a) and mean air temprature (b) in the last century. Also inverse positive correlation between precipitation (c) (December-February) and air temprature.

Temprature drops from 1909 to 1921 and from 1927 to 1942 are related to an increase in maximum snow depth. On the other hand, the temprature rise from 1942 to 1962 reflects a decrease in maximum Snow depth and winter precipitation, Small peaks in the air-tempnature record correspond to troughs in the maximum snow depth anci precipitation from 1949 to 1989.

Figure 11 shows annual variation of the three values of maximum Snow depth. winter precipitation and mean air temperalure expressed in deviations from the 100 year c1imale means.

Fig. 11. Temporal varitions of deviations from the climatic mean of maximum snow depth, precipitation and air temprature in the past century.

The maximum Snow depth has a mean value of 1.27 m with a standard deviation of 0.62 m, The precipitation mean 951.8 mm with a standard deviation of 243.5 mm The annual mean air temperature was + 1.6ଌ with a standared diviation of 1.0°C.

Discussion and Conclusions

Snow amounts in Nagaoka vary widdly from winter to winter as shown in Figure 1. Decrease of the maximum snow depth on the ground in the recent 8 years in comparison with the previous three winter of heavy snowfalls (Gosetsu) in 1984 to 1986 at Nagaoka appears to be due to the hibernal in air temperature shown in Figure 2. despite the fact that precipitation in recent winters has fact that precipitation in recent winters has decreased slightly. The maximum snow depth, various widely from year to year, but the peak maximum snow depth, derived from a satistically smoothed curve covering the last 100 years, was found to have occured in 1940. The satistically cbtained maximum value of winter precipitation was also found in 1940. Two years later, the satistically obtained minimum air temprature was found (Figs 3-5). As shown in Figure 6, variation in the winter (December-Febuary) mean temprature of thedecadally filtered values showed gradual cyclic increase in the past century with some peaks and troughs. This gendral trend is the same as reported by Reference Houghton, Jenkins and EphraumsHoughton and other (1990) obtain between 20° and 50° N in the paper, i.e a temprature rise since 1980, two troughs in 1981 and 1972, two peaks around 1960 and 1927, a trough around 1920 and a decrease from 1952 to 1940, correspond to the following specifics from Reference Houghton, Jenkins and EphraumsHoughton and others (1990): a temperature rise since 1985, two troughs in 1977 and 1989, two peaks in 1960 and 1927, a trough in 1918 and a decrease from 1952 to 1940, although it should be not that a minimum value found at Nagaoka around 1940 is not found in Houghton and others’ report. In our work, the rate of linear increase in air temperature was calculated. using 90 years’ data, as 1.35°C per centunry. The large rate of temprature rise might include the temperature rise due to city effect, because it was fOLlnd in a smaller city, Shinjo. about 250 km from Nagaoka, that the rate of temperature rise observed was 0.58°C per 100 years (Reference Nakamura and AbeNakamura and Abe, 1995 ). The rate seems to be rather large even if the negative albdo feed-back mechanism in the Northern Hemisphere (Reference Manabe and StoufferManabe and Stouffer, 1980) was involved. Temperature rise in the Northern Hemisphere should also be discussed from the point of view of the effect of the latent heat released when snow crystals are formed in the atmosphere.

It was found that there was an inverse positive correlation between the winter precipitation or the maximum snow depth and the mean winter air temperature, i.e if the air temperature decreased, the precipitation or maximum snow depth increased.

Statistically it is found th at the maximum precipitation in the last 90 years occurred in 1940. The deviation of the precipitation from the 90 year climatic mean had positive values from 1932 to 1974 to 1974. but negative values from 1905 to 1932 and from 1974 to 1994.

The statistical maximum value of the annual maximum snow depth was found in 1940, thus corresponding to the statistical maximum in precipitation. Fluctuation of the deviation of the maximum snow depth from the climatic mean did not show values as large as those associated with the fluctuation of the precipitation.

Thc minimum winter mean air temperature, observed in a 10 year moving average curve, was found in 1942. with the deviation about this mean changing from negative to positive in 1949. This change in sign may be due to global warning.

If the global temperature rise continues. it may be expected that snow in Nagaoka will decrease further, because there is an inverse positive correlation between maximum snow depth on the ground and winter air temperature in Nagaoka. But there must be a mechanism other than the usual coldness of the atmosphere when snow falls heavily. For example, as seen in Figure 7, a point marked as A corresponds to a heavy snowfall. The mechanism of heavy snowfall must be analyzed from a meterological point ofview. In addition, it has been shown statistically that Snow in Nagaoka is correlated with the “La Niña” (Reference Ferguson, Hayes, Nakamura, Ikarashi and YamadaFerguson and others. 1994). But not all heavy-snowfall years necessarily correspond to the La Niña years. Further investigation is necessary to predict the dependence of snowfall in Nagaoka on many factors, including the Asian monsoon