Design Alternatives for a Tool for Probabilistic Run-Out Calculations With MoT-Voellmy

Abstract

As part of WP 2 – Analyses of the joint Natural Hazards GBV project HARM1 (project number 20180069), it was decided to develop a tool for carrying out probabilistic runout calculations. This task was later moved to FoU Snøskred due to lack of resources. First discussions indicated that simulations with a fast, quasi-3D model like MoTVoellmy should be feasible if parallelization of the most time-consuming parts is used. Moreover, NAKSIN already contains many features that will be needed in such an application. This suggested writing a Python script that prepares the input for each simulation, then calls MoT-Voellmy, and finally counts the number of hits in each cell of the computational grid. Closer scrutiny of such a solution reveals quickly, however, that the Python script and MoT-Voellmy interact through ASCII files that must be written by the Python script and read by MoT-Voellmy. Also, MoT-Voellmy computes curvatures for the entire computational domain every time it is run, but in this application the terrain is the same for all simulations. It is therefore of interest to consider alternative approaches that might be more efficient. Below, the originally envisaged approach (Method A below) is compared to two other possible approaches B and C, and finally conclusions are drawn.