10:45 AM - 12:15 PM
[MIS07-P05] Optimal parameter distributions of depth-averaged snow avalanche model for a dense snow avalanche and its applicability
Keywords:Avalanche, Inversion, Model parameter, Numerical analysis
Although avalanches consist of solid snow and/or ice microscopically, its dynamics can be modeled as a continuum flow macroscopically. From this point of view, there have been proposed models derived with depth-averaged equations to cut-off the numerical cost (Rauter et al., 2018). Various simulators for solving such equations of motion have been proposed with different numerical solution methods, but it is necessary to set multiple model parameters in any model. However, it is difficult to give appropriate parameter values because they are empirically given, and they have not been organized. Therefore, in this study, using the method proposed by Fisher et al. (2015), we estimate the optimal parameter distribution with a certain amount of error in an observed case, and use it to reproduce another case.
As the observation, an avalanche recorded on steep slopes in Mogami-gun, Yamagata Prefecture, Japan is employed. First, based on the time-lapse images of avalanche cases and aerial photography results by UAV, the velocity of the tip and runout area of the avalanche, the thickness of the debris, etc. are obtained. Next, appropriate ranges to multiple model parameter values were given, and then randomly create a set of parameters assuming a uniform distribution. Avalanche calculations are performed using the created set of parameters and compared with the values obtained from observations. If the comparison result was above the given threshold level, it was accepted and new parameter distributions were constructed by the accepted parameters. We discuss whether the constructed parameter distributions can reproduce the dynamics of another avalanche that occurred in the same area.
Reference
Fischer, J.-T., Kofler, A., Fellin, W., Granig, M., and Kleemayr, K., (2015) Journal of Glaciology, 61, 229.
Rauter, M., Kofler, A., Huber, A. and Fellin, W. (2018) Geosci. Model Dev., 11, 2923-2939.
As the observation, an avalanche recorded on steep slopes in Mogami-gun, Yamagata Prefecture, Japan is employed. First, based on the time-lapse images of avalanche cases and aerial photography results by UAV, the velocity of the tip and runout area of the avalanche, the thickness of the debris, etc. are obtained. Next, appropriate ranges to multiple model parameter values were given, and then randomly create a set of parameters assuming a uniform distribution. Avalanche calculations are performed using the created set of parameters and compared with the values obtained from observations. If the comparison result was above the given threshold level, it was accepted and new parameter distributions were constructed by the accepted parameters. We discuss whether the constructed parameter distributions can reproduce the dynamics of another avalanche that occurred in the same area.
Reference
Fischer, J.-T., Kofler, A., Fellin, W., Granig, M., and Kleemayr, K., (2015) Journal of Glaciology, 61, 229.
Rauter, M., Kofler, A., Huber, A. and Fellin, W. (2018) Geosci. Model Dev., 11, 2923-2939.