ctsm_ppe_analysis

Analysis of the Community Terrestrial Systems Model (CTSM) perturbed parameter ensemble (PPE) output.

Startup tips

1. This repository only works on the NCAR HPC presently
2. You should clone the full repository rather than downloading individual notebooks
3. We provide a python environment that will provide the various packages referenced herein.

Creating the ppe-py environment on casper

1. log into casper via ssh
2. cd ctsm_ppe_analysis (or wherever you cloned this repo)
3. module load conda
4. conda env create -f environment.yml (this can take 15+ minutes)
5. this only needs to be performed once

Using the templates

1. log into jupyterhub.hpc.ucar.edu via a web browser

2. see quickstart info: https://www2.cisl.ucar.edu/resources/jupyterhub-ncar

3. click production and then select the Default server

4. we usually select the casper batch queue

5. I typically make the following two edits:

   Specify Memory per Node in GB (-l mem=N)

   14

   Wall Time HH:MM:SS (24-hour max)

   12:00

6. launch server

7. navigate (using the left side panel) to ctsm_ppe_analysis/notebooks and select a notebook

8. select [conda env: miniconda3-ppe-py] as your kernel (top right)

9. now you should be able to run any of the templates cell by cell