Announcing ODINN.jl, a global glacier model using Universal Differential Equations to learn empirical laws of glacier physics from large remote sensing datasets. This is my first Julia package, and the result of the last two years of work. Still pretty much a noob at this, and lots of things to improve (e.g. new and cleaner API, more tests…).
ODINN.jl is heavily based on the SciML ecosystem, using differential equation solvers from DifferentialEquations.jl, SciMLSensitivity.jl and Flux.jl for UDEs, and PyCall.jl to reuse many tools from the Open Global Glacier Model (OGGM) in Python.
We also have a new preprint describing our setup and the next steps: GMDD - Universal Differential Equations for glacier ice flow modelling
These are just the first steps, so we’d love to hear your feedback!