A new machine learning approach trained on winter and spring climate conditions offers improved forecasts of summer fire activity across the western United States.
A new project looks to unearth information about and learn from ancient underwater landslides buried deep beneath the seafloor to support New Zealand’s resilience to natural hazards.
Les images satellite permettent de mesurer les oscillations du lac de lave du Nyiragongo (RD Congo). Ces mesures renseignent sur la dynamique du volcan et aident à anticiper ses éruptions futures.
Satellite data from Nyiragongo Volcano, Democratic Republic of Congo, track changes in summit-crater lava levels that provide a window into eruption dynamics and aid in forecasting future activity.
How big might future volcanic eruptions be? Crystals carry information to answer this and machine learning methods can visualize and interpret this multidimensional data.
Waveform inversion of regional earthquakes reveals velocity anomalies interpreted as subducting seamounts that control an enigmatic segmentation in plate coupling along the Hikurangi margin.
A simple, yet quantitative, index is demonstrated to quantify reductions in the peak flood resulting from multiple reservoirs, arranged in series along the same river reach.
New methods for identifying debris flow-shaped channels improve hazard quantification and highlight how high uplift rates and fractured bedrock facilitate debris flow-dominated landscape evolution.
A new initiative is bringing together scientists to address fundamental questions about subduction zone geohazards, using the latest advances in observation technology and computational resources.
Using susceptibility models to forecast the potential locations of landslides is a key tool in mitigating landslide hazard, but are existing approaches appropriate in dynamic mountainous settings?