glacier_prethicktor

A machine learning approach to predicting glacier thicknesses.

We treat estimating glacier thickness as a regression problem as opposed to ice velocity flow models. We train machine learning models with Glacier Thickness Database thickness measurements as an independent variable co-registered with surface attributes from the Randolph Glacier Inventory as dependent variables to estimate thickness.

Requirements

• python 3.8.10
• tensorflow v. 2.12.0
• Jupyter

Install conda environment and dependencies with the following code snippet

conda create -n glacierml python=3.8.10
conda activate glacierml
pip install ipykernel
python -m ipykernel install --user --name=glacierml --display-name="glacierml (Python3.8.10)"

Run the notebooks in number order. Just set the home path and everything should be automated.