Home

The Instructed Glacier Model (IGM) simulates the ice dynamics, surface mass balance, and its coupling through mass conservation to predict the evolution of glaciers and icefields. The specificity of IGM is that it models the ice flow by a neural network, which is trained with ice flow physical models. Doing so permits to speed-up and facilitate considerably the implementation of the forward model and the inverse model required to assimilate data.

IGM consists of an open-source Python package, which runs across both CPU and GPU and deals with 2D gridded input and output NetCDF data. Together with a companion library of ice flow emulators, IGM permits highly efficient and mechanically state-of-the-art modeling. The structure of IGM facilitates the customization of model components to applications and the coupling to other models. The computational efficiency of IGM is the result of fully-vectorized mathematical operations via the TensorFlow library.

How to start: If you don't know anything about glacier evolution modeling, you may watch first this video. In any case, I recommend starting by running IGM examples: If you are unfamiliar with python, I can do that with online Colab notebooks. Otherwise, I recommend first installing an igm python environment on your system and starting with examples. Then, you may explore step-by-step the IGM workflow, before going further into coding and custumizing IGM to your application.

Operating System: IGM was developed in a Linux environment but should be working on any system. At the time of writing, IGM worked well on Linux and Windows OS, but issues were reported for the installation of TensorFlow on mac OS.