Avalanche Hazards in Khibiny Massif, KOLA, and the new Nansen Seismograph Station

Pavel Chernouss, Yury Zuzin, Evgeny Mokrov ¹⁾, Gennady Kalabin ²⁾, Yury Fedorenko ^2,3), Eystein Husebye ^3)
1) Center of Avalanche Safety of "Apatit" JSC, Murmansk Region, Russia, ²⁾ Kola Science Centre RAS, Russia, ³⁾ Institute Solid Earth Physics, University of Bergen, Norway

Figure 1. Topographic map of the Khilbini Massif (elevation step 100m) including locations of major towns, operative mines and naturally the new 3-component Nansen station. NKK is close to several mines and besides offer an unique opportunity to study 3D wavefield responses of rough topography (Hestholm and Ruud, 1998)

Background
The Khibiny Mountains in Kola is highly snow avalanche prone during winter with hundreds of avalanches taking place each year. In the pioneering days of mining operations in the Khibiny, which commenced in 1929, the avalanche hazards were largely ignored until a tragic accident on December 5th, 1935 when 88 miners perished near Kirovsk (Figure 1). Avalanche safety measures had already been initiated in 1933, and even at this early stage artificial avalanche release experiments were conducted. The 1935 tragedy triggered more comprehensive safety measures, and various research programs for measurements bearing on physical snow conditions, precipitation, prevailing winds and other meteorological parameters. The ultimate goals of these efforts, dating back to 1935, are enhanced safety measures and avalanche forecasting. The latter goal has proved rather elusive because the avalanche releases are a nonlinear process. Nevertheless, it remains highly relevant today, as skiers 'invade' the Khibiny in winter. In the future, avalanche hazard mitigation will become increasingly important as more skiers and tourists visit the beautiful Khibiny Massif area. The project introduced here aims at physical avalanche modeling with the overall goal of risk mitigation through improved avalanche forecasting.

Figure 2. Artificial avalanche release by mortar firings into the hanging snow wall at the mountain top. The pockmarks indicate the explosion points.

Seismic loading artificial avalanche releases
Artificial avalanches can be triggered using dynamite or firing mortar rounds into the 'snow hanging wall'. These techniques are well proven measures for mitigating such hazards. Avalanches may also be released by seismic loading, which in the case of Khibiny, are caused by open pit and underground mine explosions. Over the period 1959­1999, approximately 225 avalanches were triggered in this manner, excluding those released intentionally by in situ shootings. (Figure 2) Another example is the large explosion (approximately 100 tons of dynamite) on January 20, 1998, which triggered an avalanche of 40,000 cubic meters directly into the open pit mine Centralnyi. As a result a stretch of a road 600 meters long was buried under thick snow. Even moderate pit mine explosions with charges in the range 10­100 kilograms of dynamite are observed to trigger avalanches at ranges of 2­3 km away from the source, while the larger 10­100 tones explosions can trigger avalanches at least 15 km away.

The Nansen station seismic loading monitoring
As previously mentioned, observations indicate that a causal relationship exists between seismic loading and avalanche releases. To model the phenomena, we require a more quantitative relationship, which, in turn, motivated our deployment of the Nansen 3-component station in Khibiny, near the Kirovsk mining town (see Figure 1). The station became operational January 5, 1999 and over 5 months, hundreds of mining explosions were recorded. A few recordings from this station are shown in Figure 3. The SP seismometers used ground velocities which were converted to peak ground accelerations (PGAs) and similar measures. The real challenge is to simulate PGAs for an arbitrary explosion over larger Khibiny areas using 3D wave field synthetics, and properly accounting for topographic focusing/defocusing effects (Hestholm and Ruud, 1998).

a)	b)
Figure 3. Nansen station seismograms from a) a local mine explosion (18 km away) and b) a teleseismic event. The presumed PmP-phase in a) illustrate the importance of topography in wavefield modeling. The phase polarization is elliptical, and the amplitude is almost the same at each of the 3 components.

Avalanche modeling and forecasting
Several attempts of using avalanche data for statistical forecasting for small areas of the Khibiny have been under taken, but apparently without much success (Chernous and Fedorenko, 1998). Most successful were forecasts of 'avalanche situations' during periods of heavy snow accumulation, implying that some avalanches did occur after delays of 2­4 days not entirely unexpected. The avalanche database (dating back to 1933) has been subjected to various kinds of multivariate statistical analysis, but it remains somewhat unclear which physical parameters are most diagnostic for avalanche releases. Parallel to these investigations, studies on stochastic 3D models simulating avalanche releases have commenced (Chernouss and Fedorenko, 1998). The data from the Nansen station would be most useful, namely simulating the exceedance (seismic loading) of critical friction force limits in order to initiate an avalanche release.

Perspectives
Avalanche safety measures are well handled by the Apartity Avalanche Safety Centre of JSC, Kirovsk, but areas of operations are limited to mining towns and their surroundings. Our project aims at avalanche forecasting for larger areas, in particular those popular with weekend skiers. The most difficult part of the project would be to merge wind modeling and snow accumulations with changing snow conditions and avalanche triggering levels into an Avalanche Hazard Model. Our approach would, in some respects, be similar to those used in earthquake risk analysis and prediction. Even if we are only moderately successful, more stations would be deployed for monitoring avalanche occurrences in the Khibiny Massif.

We invite any interested IRIS members to join us in our avalanche monitoring and modeling efforts.

Acknowledgments
We express our sincere thanks to the Nansen Foundation, Norwegian Academy of Sciences and Letters, Oslo for the financial support that enabled us to deploy and operate the Nansen station.

References
P. Chernouss and Yu. Fedorenko, 1998. Probabilistic Evaluation of Snow Slab Stability on Mountain Slopes. Annals Glaciology, 26, 303-306.

Hestholm. S and B. O. Ruud, 1998. 3-D finite difference elastic waves modeling including surface topography.Geophysics,63, 613-622