A blueprint for an Indonesian Landslide Early Warning System (BILEWS)

BILEWS focuses on landslides caused by heavy rainfall to improve village-level warnings in Indonesia. The goal was to use physical models combining rainfall patterns, soil and landscape features to better predict landslide locations. Despite insights into local causes, data uncertain is too large for predictions and collecting more detailed data is impossible on a national scale. Neural networks (AI) do seem suitable for predicting landslide zones. This would fit better in the warning system, which is embedded in laws and regulations, so major methodological changes are not feasible. Productive meetings with stakeholders gave better understanding of early warning in reality.