Earthquakes claimed around 750,000 human lives in the two decades 1998–2017. Along with building collapse and tsunamic flooding, earthquake-triggered landslides are a major cause of fatalities. For instance, the Mw 7.9 Wenchuan Earthquake (China) caused around 200,000 landslides that killed more than 20,000 people.
GFÚ researcher, John Jansen (Surface Processes & Palaeoclimate) is a co-author among a team of Chinese researchers who this week (9 May) published in the National Science Review an important advance in predicting earthquake-triggered landslides. The study employs an advanced deep-learning model coupled to a new global database of 400,000 landslides associated with 38 of the most catastrophic earthquakes over the past 50 years. The model demonstrates the capacity to predict the probability of landsliding for any earthquake worldwide with an average spatial accuracy of ~82 % in less than a minute. This represents a transformative step in geohazard prediction.
The study can be accessed here: https://doi.org/10.1093/nsr/nwaf179
Photo: The giant 250 m high palaeolandslide at Diexi near the eastern margin of the Tibetan Plateau. Part of this landslide was remobilised during the catastrophic 1933 Sichuan earthquake.