Skip to main content

Forecasting volcanic activity using deep learning

Project description

Forecasting the threat of destruction

There are an estimated 1 500 potentially active volcanoes on land worldwide. Many appear to have been ‘dead’ for centuries, but may be only ‘sleeping’. In 2019, there were 73 confirmed eruptions from 70 different volcanoes. While deadly eruptions are not common, they do happen. This is why forecasting eruptions – the time and severity – is so important. The EU-funded DEEPVOLC is applying advances in artificial intelligence and satellite monitoring to create transformative new geodetic datasets. Specifically, it will base observations on the recently-launched European satellite mission, Sentinel-1S, which has advanced scientists’ ability to measure surface deformation at all volcanos worldwide. Ultimately, it will create a system to monitor and forecast deformation, which is caused by magma migrating beneath – an indicator of potential volcanic activity.

Field of science

  • /natural sciences/computer and information sciences/artificial intelligence
  • /natural sciences/earth and related environmental sciences/geology/volcanology
  • /natural sciences/computer and information sciences/artificial intelligence/machine learning/deep learning

Call for proposal

ERC-2019-COG
See other projects for this call

Funding Scheme

ERC-COG - Consolidator Grant

Host institution

UNIVERSITY OF LEEDS
Address
Woodhouse Lane
LS2 9JT Leeds
United Kingdom
Activity type
Higher or Secondary Education Establishments
EU contribution
€ 1 941 270

Beneficiaries (2)

UNIVERSITY OF LEEDS
United Kingdom
EU contribution
€ 1 941 270
Address
Woodhouse Lane
LS2 9JT Leeds
Activity type
Higher or Secondary Education Establishments
UNIVERSITY OF BRISTOL
United Kingdom
EU contribution
€ 58 225
Address
Beacon House Queens Road
BS8 1QU Bristol
Activity type
Higher or Secondary Education Establishments