An Empirical-Statistical Model for Landslide Runout Distance Prediction in Indonesia

Abstract

There have been many attempts and methods for predicting landslide-affected areas; empirical methods, numerical methods, and laboratory models are commonly used for prediction. Laboratory and numerical models require an input of parameters that are difficult to determine accurately. At the same time, empirical statistical methods use statistical methods based on historical data of landslide events to form an empirical model. Statistical analysis of empirical observations builds a possible relationship between disaster area characteristics and slide behavior because it does not require detailed mechanics of avalanche movement; the empirical-statistical model is a simple and practical tool in the initial assessment to predict the sliding distance of an avalanche that will occur. The main discussion of this study is that the volume of avalanches (V) has a more significant influence than the height of the slope (H) on the length of the avalanche (L) that occurs. Fifty-nine data on landslide events that have occurred in Indonesia are used to a prediction model for landslide events reviewing the slope geometry parameters in the form of H, slope (θ), and V and discussing the main factors that affect the sliding distance of avalanches that have not been discussed in research in the Indonesian territory. The analysis shows that H has a significant effect on the sliding distance of the avalanche compared to V. The best model produced to predict the sliding distance of the avalanche is L = 6.918 H^0,840 and produces an average error rate of 29% for the landslide measurement data.