Skip to content

Teaching with IGM

IGM runs a 3D ice-flow model on any glacier of the world from a short configuration file, which makes it a good tool to teach glacier modelling. With oggm_shop, a glacier is set up from its RGI ID alone; students can then calibrate the model, and run its past and future evolution in a few minutes.

Glacier Modelling, a Master course at the University of Lausanne, is built on IGM, and all its material is open. Each student picks a glacier by its RGI ID and, step by step in a notebook, calibrates the ice flow (sliding and ice softness) against observed speeds, calibrates the surface mass balance so that the glacier reaches its Little Ice Age moraines and its present extent, then runs it into the future under climate scenarios and estimates its response time. The notebooks run online on any Jupyter service (Colab, Binder, …) or locally.

Light emulator for teaching

Choose the pretrained emulator dahunet_edu: it is very small and runs fast on a laptop CPU. It is less accurate than the standard emulators, but good enough for students to explore.

processes:
  iceflow:
    numerics:
      Nz: 2
      basis_horizontal: q1
      basis_vertical: molho
    unified:
      nbit_init: 0
      retrain_freq: 0
      network:
        pretrained: true
        pretrained_path: dahunet_edu.keras