OUTLINE AND CHARACTERISTICS OF SNOW DISASTER DURING 2005_2006 WINTER SEASON IN JAPAN.pdf

Toru TAKAHASHI

Chiba University, Japan

Tsukasa TOMABECHI

Hokkaido Institute of Technology, Japan

Seiji KAMIMURA

Nagaoka University of Technology, Japan

OUTLINE AND CHARACTERISTICS OF SNOW DISASTER

DURING 2005-2006 WINTER SEASON IN JAPAN

ZARYS I CHARAKTERYSTYKA KATASTROF

SPOWODOWANYCH Ś NIEGIEM PODCZAS ZIMY 2005-2006

W JAPONII

Abstract: We had heavy snow during 2005-2006 winter season in Japan. Totally, 152 people were killed by

snow, and 902 people were heavily insured. On the other hand, the number of collapsed residential buildings was

only 18, and 28 buildings were half damaged. In this paper, the outline of the disaster was pointed out and the

characteristics and estimated reasons were discussed.

Streszczenie: Podczas cięŜkiej zimy w Japonii w sezonie 2005-2006 ogółem 152 osoby zostały zabite, a 902

cięŜko ranne na skutek katastrof spowodowanych przez śnieg. Z drugiej strony liczba zniszczonych budynków

mieszkalnych wyniosła jedynie 18, a 28 zostało w połowie zniszczonych. W referacie przedstawiono zarys ka-

tastrof, a charakterystyczne i przewidywane powody omówiono.

1. Introduction

The Meteorological Agency of Japan named the heavy snow during 2005-2006 winter season

in Japan as “Heavy snow disaster in 18th. year of Heisei period (2006).” It was he first time

after heavy snow in 1963. The number of death was 152. That was the second largest number

during recent 60 years. The regional distribution of death is shown in Figure 1. We usually

have heavy snow on the Sea of Japan side. However, even in the Pacific Ocean side, there were

some incidents and people were killed by snow. In addition, 902 people were heavily insured

and 1243 were slightly insured.

331

The Sea of Japan

The Pacific Ocean

Figure 1. Regional distribution of death in the heavy snow during 2005-2006 season.

On the other hand, it could be said that the number of damaged buildings was very small.

Totally, 18 residential buildings were collapsed, 28 were heavy damaged and 4667 were

slightly damaged. The numbers were quite small comparing with those of 1963 or 1981. The

comparison of disasters among 1963, 1981 and 2006 is shown in Table 1

In this paper, the past records of snow depth and snow accumulation in short terms are

compared and discussed statistically. Then, the reason of human loss is considered using the

data of snow accumulation in short term.

Table 1. Comparison of typical snow disasters in Japan

human

residential buildings

non-residential

buildings

year

heavy

damaged

slightly

damaged

dead

missing

injured collapsed

public

other

1963

228

356

753

982

N/A

1981

133

2,158

165

301

N/A

2006

152

2,145

4667

145

2,314

2. Comparison with past heavy snow

Figure 2 shows transition of annual maximum snow depth that standardized by the value for

mean recurrence interval of 100 years for 17 observation points in Japan. This figure shows

snow in the 2005-2006 season was not so “heavy” snow comparing with the past heavy snow.

Then, it was the question that why such a many people were killed by snow?

332

Figure 2. Transition of annual maximum snow depth at 17 observation points in Japan.

Figures 3 and 4 show the statistical analysis for annual maximum snow depth and annual

maximum snow accumulation in 7 days, respectively. The annual maximum values are plotted

on the Gumbel probability paper by Hazen plotting rule that is expressed as Equation (1). This

plotting rule is suitable with parameters given by Gumbel’s moment method.

( ) = 1 -

2 i

(1)

2 N

where, F X ( x i ) is non-excessive probability for i th largest value and N is the total number of data.

From these figures, snow fall in 2006 season could not be said as “extremely heavy snow”

except of Tsunan, where is mountain side town in Niigata prefecture.

Figure 3. Statistical analysis for annual maximum snow depth at three observation points

in Niigata Prefecture .

Figure 4. Statistical analysis for annual maximum snow accumulation in 7 days

at the same observation points with figure 3.

333

F X x i

Figure 4 shows statistical analysis for annual maximum snow accumulation in 7 days that

is defined as shown in Figure 5 (Izumi, Mihashi & Takahashi 1988). The values observed in

2006 do not seem to be “extreme value” except of Tsunan. Then, the authors investigated their

transition process. Figure 6 shows comparison of plotting rules.

Figure 5. Definition of annual maximum snow accumulation in n days (Izumi et al. 1988).

Figure 6. Comparison of typical plotting rules and linear regressions.

Figure 7 shows time series of snow depth and snow accumulation in 3 or 7 days at Tsunan

in 2006, 1976 and 1981 season. In 2006 season, the value of snow accumulation in 7 days

(SI-7) indicates periodic heavy snowfall especially in December. In 1976, the peak value of

SI-7 indicated almost the same value with 2006, but it was not continuous. In 1981, the value of

snow depth was larger than 2006, but the value of SI-7 was not so large as that of 2006.

Figure 7. Transition process of snow depth and snow accumulation in 3 and 7 days (Tsunan).

334

Figure 8. Transition process of snow depth and snow accumulation in 3 and 7 days

at Nagaoka, Tokamachi and Aomori in 2005-2006 winter season.

As shown in Figure 8, the patterns of snowfall in 2005-2006 season are similar even if the

region is different from Tsunan. Therefore, the authors suppose the time series characteristic of

snowfall in 2005-2006 season is one of the most appropriate reasons why many people were

killed by snow.

3. Standard, building code and failure

The latest version of Japanese building code (2000) is following AIJ (Architectural Institute of

Japan) standards: Recommendations for Loads on Buildings (1993). In the recommendation,

snow loads were estimated based on ground snow depth and equivalent snow density. The

relation between snow depth and snow density is shown in Figure 9. When the construction site

is far from observatories, snow depth may be estimated from altitude and sea ratio that is de-

fined as shown in Figure 10. In the case, snow depth is estimated as follows,

d = a × altitude + b × sea ratio + g

(2)

are defined from multiple regression analysis for

each zone. The coefficients are listed in the bulletin of the ministry of construction (in this paper,

the concrete values are neglected.) The concrete values for design snow load are provided by

each local government in their ordinance. In there, isopleth maps and zone maps are both used

depending on the local government. Basically, the coefficients for equation (2) were estimated

for 50-year mean recurrence interval (MRI) value. Therefore, most of all buildings that de-

signed by allowable stress design adopt 50-year MRI value for snow load. On the other hand,

AIJ recommendation recommends 100-year MRI value for basic value for design because its

probability of exceedance in 50 year is similar to average value of 50-year maximum value.

This comparison can be seen in Figure 6. The largest value plotted by Gringorten plot is placed

on 89.5 year that is actually equal to the average value of 50 year maximum value when the

distribution is Gumbel distribution. The formula is expressed as follows:

and

( ) = 1 -

0.44

0.12

(3)

335

where, d is ground snow depth,

F X x i

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: