научная статья по теме AN EXPERIENCE OF AVALANCHE RISK MAPPING FOR SMALL MOUNTAIN AREA Геофизика

Текст научной статьи на тему «AN EXPERIENCE OF AVALANCHE RISK MAPPING FOR SMALL MOUNTAIN AREA»

УДК 551.578

An experience of avalanche risk mapping for small mountain area

© 2011 г. E. Semakova1, S. Myagkov1, R. Armstrong2, A. Racoviteanu2

1Hydrometeorological Research Institute, Uzhydromet, Tashkent, Uzbekistan; 2University of Colorado, National Snow and Ice Data Center, Boulder, Colorado, U.S.A.

ella9sem@gmail.com

Статья принята к печати 20 января 2011 г.

Avalanche regime, climatic indexes, snow depth, topography variables.

Лавинный режим, толщина снега, топографические и климатические характеристики.

The large-scale mapping technique for snow depth with different probability is considered on example of Kam-chik pass region. Input data were digital elevation model, meteorological variables (precipitation and air temperature) and data on snow depth measured by remote stakes in the avalanche catchments. The climatic characteristics that define snow and avalanche regime for given area were identified.

На примере района перевала Камчик рассмотрена методика построения крупномасштабных карт толщины снега разной обеспеченности. В качестве исходных данных использована цифровая модель рельефа, данные метеорологических наблюдений (осадки и температура воздуха), измерений толщины снега по дистанционным рейкам. Выявлены индикационные показатели климата, определяющие снеголавинный режим в данном районе.

Introduction

Snow avalanches are serious natural hazards for roads in the region of Kamchik pass that is the part of Tashkent-Osh highway worldwide known as «Great Silk Road». This road is of great importance for the country because it connects the two most densely populated parts of Uzbekistan. Maps of snow cover depth are useful for the snow avalanche forecasts, avalanche risk assessment and designing of avalanche defense structures in this region. For example, the maximum snow depth with 5% probability map is needed to define the highest snow loads on anti-avalanche constructions. They are also important to estimate future avalanche risk under possible climate change.

The technique of large-scale mapping for snow depth with different probability was developed for Chimgan valley - the famous recreation zone of Uzbekistan [3]. This technique was adapted for specific conditions of Kamchik region. There are only some expert assessments for prediction of avalanche release on the old road. The situation is changed while the new road and anti-avalanche constructions were constructed for last ten-year. The difficulties of analysis of snow-meteorological characteristics and avalanche regime through this road did not allow until now to reveal the regularities of spatial pattern for the snow depth and to performance the snow depth, avalanche formation and run-out zones mapping, which are necessary components for avalanche risk assessment. The given work is devoted to issues of snow depth distribution studies and the searching of indicative climatic characteristics for definition the avalanche regime in this region.

Avalanche regime

Kamchik avalanche recording station located at the elevation of 2145 m above sea level on the North part of Kuramin ridge (Western Tien Shan) can be used for description of corresponding climatic conditions. As a station with continuous meteorological and snow avalanche observations, it started to collect data in 1982. Episodic stationary observations were from 1964. Detailed terrain and climatic description is presented in [1]. The dominate avalanche type is direct action avalanches (storm-packed snow avalanches) (Fig. 1). They release during or after snowfalls as a result of an overloading of slopes by newer snow and during a storm. The precipitation is accompanied by North

Fig. 1. Avalanche amount distribution for observation period in the Kamchik pass region:

a - by genesis snow content: 1 - fresh snow, 2 - direct action (storm snow), 3 - old snow, 4 - other, 5 - not determined; b - by moisture snow content: 1 - loose (dry), 2 - wet Рис. 1. Распределение количества лавин за период наблюдений в районе перевала Камчик:

а - по генезису снега: 1 - свежевыпавшего, 2 - метелевого, 3 - старого, 4 - прочие, 5 - не установлено; b - по влажности снега: 1 - сухие, 2 - мокрые

- 67 -

5*

winds which results in increase of snow on leeward slopes, where snow depth can reach 4-5 m. The strong wind causes redistribution of snow. Moving snow results in snow cornices and drifts formation. The average wind velocity varies from 2 up to 8 m/s; the instantaneous velocity during snowstorms can reach 25-28 m/s. The avalanche prone season is NovemberApril, as a rule. Its average long-term duration is 101 days. Volume of avalanches ranges from several cubic meters up to 500 000 m3, but mostly avalanches have volume less than 5000 m3.

Data and investigation methods

Kamchik station was assumed as the base point for snow depth definition with different probability. For prolonging these data the equations of relationships for daily values on air temperature (coefficient of correlation is R = 0.86) and ten days precipitation sum (R = 0.68) between Kamchik and Tashkent stations were derived as well as relation for snow depth on meteorological sites between Kamchik and Chimgan stations (R = 0.81). Calculation of everyday snow depth on meteorological site, since November 1, was made with use of calculations of snow increment and decrement, and also the daily solid precipitation sum by approach for Chimgan valley [3]. For daily solidphase precipitation sum calculation we used the method of Glazyrin [2]. The simulation results revealed that ratio of calculation between standard error (S) and standard deviation (a) of snow depth daily values is varied from 0.09 to 0.42. It corresponds to correlation coefficients (R) from 0.9 to 1.0.

The more special attention is paid to revealing of snow cover relationships between investigated avalanche catchments and meteorological station area taking into account the elevation difference, exposition of slopes and surface type. Input cartographic information was the digitized map on this region (1:25 000 scale) with an area of 36.5 km2. The ArcGIS 9.3 was applied to generate the digital elevation model (DEM) using the TOPOGRID algorithm and such layers as elevation isolines, streams and elevation tops. It was revealed that this DEM has better quality than the ASTER DEM. Snow data at the dates of maximum snow accumulation during winters with different snow amount have been collected. Snow depth was measured by 38 remote snow depth-gauges. Searching the snow depth distribution features we are operating with topography variables such as: elevation, terrain aspect, terrain slope, plan and tangential curvature, distance to thalweg and distance to ridge and etc.

The selection of relief parameters was carried out by the mean the 7 snow surveys of remote snow depth-gauges (February-March 1984). The most important factors that affect snow depth distribution are G-dis-

tance to ridge, C-curvature and W - so called wind index:

_ [cos((o-59.9)sin(a+28.1),ifcx*0; W = \

0, if a = 0,

where is ro - Terrain Aspect (degree) and a - Terrain Slope (degree), the phases were found by Microsoft Excel «Solver Parameters». The Set Target Cell is maximum of correlation coefficient between snow depth and W. The wind index expresses the effect of snow mass accumulation on leeward slopes (cosine is positive) and its removing from windward (cosine is negative). This process becomes more increased for steeper slopes (it takes into account a second multiplier).

The theory of the mean curvature of a surface is the basis for C calculation. The plan form curvature influences convergence and divergence of flow as it presents the surface form which is perpendicular to the slope direction. The calculation algorithms of G, C, ro and a characteristics are shown in ArcGIS Help [4].

Comparison of actual snow depth with accounted snow depth by Distance to Ridge, Curvature and Wind index revealed the following assessments: the multiple correlation coefficient is 0.83, the corresponding coefficient of determination is 0.68, the value of ^-criterion is 18.81, the corresponding significance level p is about 0 and the standard error is 20.5 cm.

The coefficients (k) for following equation were derived by least squares procedure with taking account of mean snow depth for February and March 1984, 1987 and 1991

hi

-J- = kGi+k7Ci + L W, + k.

U 1 J J J

ms

where hj is the snow depth for slopes and hms is the snow depth for meteorological station. The accuracy of derived equation is following: standard error 35 sm, S/a = 0.68 on hydrological tolerance is 0.8; the multiple correlation coefficient is 0.72, the factors explain the resulting feature by 0.52%, the ^-criterion is 12.4 and p is certainly less than 0.05. Thus, this model is statistically significant. The test for independent series showed that the algorithm accuracy is not always satisfactory by definite date. Occasionally, snow depth series are variable from year to year. Mispositioning of remote snow depth-gauges on slopes can probably cause it. But there is a good result of testing for mean snow depth for February and March in years of maximal snow accumulation (S/a < 0.80). It allows us to use this equation for assessment of maximal snow depth on the slopes in this region for year with definite probability and to consider the changing of this value with relation to climatic conditions.

The prolonging of meteorological data on meteorological site of Kamchik station allowed calculating the maxi-

E. Semakova et al.

Fig. 2. The maximal snow depth map for winter season with 50% probability: 1 - Kamchik snow avalanche station; 2 - snow depth-gauges; 3 - highway; 4 - rivers Рис. 2. Карта максимальной за сезон толщины снега 50%-й обеспеченности: 1 - снеголавинная станция Камчик; 2 - снегомерные рейки; 3 - дорога; 4 - реки

Climatic Indexes affected on snow avalanche regime in the Kamchik region

Snow avalanche characteristics Climatic Indexes R

Maxima

Для дальнейшего прочтения статьи необходимо приобрести полный текст. Статьи высылаются в формате PDF на указанную при оплате почту. Время доставки составляет менее 10 минут. Стоимость одной статьи — 150 рублей.

Показать целиком