Global Landslide Nowcast: Daily Probabilistic Hazard Maps

Global maps provide daily, near-real-time probabilistic assessments of rainfall-triggered landslide hazard at a 1 km x 1 km resolution. The Landslide Hazard Assessment for Situational Awareness (LHASA) version 2.0 model uses machine learning to integrate satellite precipitation data from GPM, soil moisture from SMAP, and other environmental factors. This archive contains a retrospective daily record, with a minimum data latency of 5 hours for operational nowcasting.

Use Cases

• Issuing landslide early warnings based on probabilistic hazard thresholds.
• Validating landslide prediction models using the machine learning-based LHASA 2.0 framework.
• Analyzing regional landslide risk trends from the daily, global 1 km resolution record.
• Integrating satellite-derived soil moisture and precipitation data for hydrological hazard assessment.

Strengths

• Model employs a machine learning framework (LHASA 2.0) for improved probabilistic output over a prior heuristic version.
• Provides global, daily coverage at a 1 km spatial resolution.
• Integrates multiple satellite data sources (GPM precipitation and SMAP soil moisture) for a multi-factor hazard assessment.
• Designed for near-real-time situational awareness with a stated operational latency of 5 hours.

Limitations

• Key metadata conflicts: last update date is listed as both 2021-02-10 and 2026-04-10 across sources.
• Specific column names and dataset size (row count, file size) are not provided in any source.
• The archive is a daily retrospective product, while the operational model could run every half-hour.

Provenance

Source

GES DISC (NASA)

Collection Method

Satellite-derived modeling using the Landslide Hazard Assessment for Situational Awareness (LHASA) version 2.0.

Freshness

2026-04-10 02:26:13.354085

Geography

Global