CSDMS 2023 Webinars

IGM, a data assimilation and glacier evolution model boosted by deep-learning

Registration link:

Guillaume Jouvet

Inst of Earth Surface Dynamics, University of Lausanne, Switzerland

Deep-learning emulators permit to reduce dramatically the computational times for solving physical models. Trained from a state-of-the-art high-order ice flow model, the Instructed Glacier Model (IGM, is an easy-to-use python code based on the Tensorflow library that can simulate the 3D evolution of glaciers several orders of magnitude faster than the instructor model with minor loss of accuracy. Switching to Graphics Processing Unit (GPU) permits additional significant speed-ups, especially when modeling large-scale glacier networks and/or high spatial resolutions. Taking advantage of GPUs, IGM can also track a massive amount of particles moving within the ice flow, opening new perspectives for modeling debris transportation of any size (e.g., erratic boulders). Here I give an overview of IGM, illustrate its potential to simulate paleo and future glacier evolution in the Alps together with particle tracking applications, and do a quick live demo of the model.

Please acknowledge the original contributors when you are using this material. If there are any copyright issues, please let us know ( and we will respond as soon as possible.

Of interest for:
  • Marine Working Group
  • Terrestrial Working Group
  • Coastal Working Group
  • Hydrology Focus Research Group
  • Critical Zone Focus Research Group
  • Human Dimensions Focus Research Group
  • Geodynamics Focus Research Group
  • Ecosystem Dynamics Focus Research Group
  • Artificial Intelligence & Machine Learning Initiative
  • River Network Modeling Initiative