r.slope.stability contains built-in functions for validating and visualizing the simulation results. These functions enable a rapid evaluation of the model outcomes without the need for manual post processing. Explore the results by viewing maps and plots derived from selected model runs.
The simulation results obtained for shallow landslides with a preliminary version of Release 20140425 are illustrated, based on a pixel size of 5 m. Note that the results shown are not necessarily the most up to date ones, but are visualized for demonstrating the functionalities of r.slope.stability only. Selected simulations for the Collazzone Area are discussed in detail by
The maps below show the factor of safety and the slope failure probability obtained with the soil class mode of r.slope.stability. Further, additional output maps optionally produced along with the factor of safety are illustrated. The plots on the right side relate the model results to the observed shallow landslide areas. For the factor of safety, the four-fold radar chart illustrates the rates of true and false positive and negative predictions. Depending on these rates, the prediction may be classified as successful, unsuccessful, conservative or non-conservative. Here, the plot indicates a fairly successful conservative prediction. For slope failure probability, the ROC Curve tests the prediction rates at different threshold levels. A straight line (AROC=0.5) means a failed prediction, AROC=1.0 means a perfect prediction. Here, the prediction may be classified as fair.