Parallelized Greenland Ice Sheet Basin Data Query

A Python workflow for extracting Modèle Atmosphérique Régional (MAR) variables over Greenland Ice Sheet (GrIS) drainage basins and computing basin area-weighted time series of these variables. This workflow supports MPI parallelization in high-performance computing (HPC) systems.

Features

• Extracts MAR variables at user-defined locations on the Greenland Ice Sheet
• Clips MAR data to ice sheet basins
• Computes area-weighted basin averages of MAR variables
• Parallel execution via MPI

Requirements

• Python 3.9+
• mpi4py
• os
• pandas
• geopandas
• numpy
• shapely
• pathlib
• xarray
• rioxarray

These dependencies can be found in the cee690_tf.yml file.

The following steps walk through how to install and run this project on an HPC system:

1. Clone this repository.

git clone https://github.com/mlmac-seid/GrIS-basin-query.git
cd GrIS-basin-query

2. Request an interactive node (on a Slurm-based HPC system).

srun --pty -n 4 --mem=64G bash -i

3. Create and activate the virtual environment.

conda env create -f cee690_tf.yml
conda activate cee690_tf

4. Install the package as editable.

pip install -e .

5. Run the package.

gris-basin-query

Input File Formats

Coordinates csv:
Must include columns:

• latitude, longitude
  or
  
• lat, lon
  

Years csv:
Must include column:

• year

MAR variables csv:
Must include column:

• variable
  or
  
• var
  or
  
• variables
  or
  
• vars