seb-1/enso_workshop

Scripts for the ENSO/SMILE workshop August 2021

ENSO_workshop

Scripts for the ENSO/SMILE workshop August 2021

Use https://jupyterhub.hpc.ucar.edu to run the notebooks and select the Pangeo Kernel.

Examples include simple steps like loading large ensemble output, doing simple calculations, and plotting.

data on Cheyenne

MMLEA: /glade/collections/cdg/data/CLIVAR_LE

CMIP6 SMILEs: /glade/scratch/rwills/cmip6_ensembles

Prerocessed GSAT and Nino3.4 SST for some models: /glade/scratch/milinski/SMILEs

custom environment

(optional, more flexibility)

To use a custom conda environment on CISL systems that allows installing your own packages, such as python cdo bindings, do the following:

Download the conda installer into your home directory: wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh

and run bash Miniconda3-latest-Linux-x86_64.sh

Create a new environment: conda create --name jlab_37 python=3.7 jupyterlab matplotlib cartopy cdo netcdf4 numpy pandas python-cdo scipy seaborn tqdm xarray ipykernel dask

Note that python >3.7 can cause problems with jupyterhub at CISL. ipykernel is required so that jupyterhub can use the environment.

Consider moving the environments (which can be ~4GB) to a different location. This can be done by creating the file ~/.condarc with the following content:

envs_dirs:
  - /glade/work/USER/conda/conda-envs
pkgs_dirs:
  - /glade/scratch/USER/conda/conda-pkgs
changeps1: false
auto_activate_base: false
channels:
  - conda-forge
  - defaults