GFDL Notebooks

(Previously MAR - Model Analysis Repository)

The Latin word "mar" translates to "sea". This repository will contain a collection of (mainly) ocean-focused analyses to inform next-generation ocean and climate model development.

Ways to Run MAR

1. Interactively (clone the repository, edit the notebooks, and run)
2. Execute the batch script run_mar.sh
3. Visit https://dora.gfdl.noaa.gov/analysis/mar

Contributing to MAR

Jupyter notebooks are the encapsulation of a particular analysis. There are relatively few constraints on how an analysis built, but there are a few interfaces to be aware of:

Configuration / Environment Variables

The batch and web engines for MAR (items 2 and 3 above) will set two runtime environment variables. Use one or both of these fields to determine the top-level path to a model experiment to analyze:

• MAR_DORA_ID: The experiment ID in the dora database
• MAR_PATHPP: The top-level path to the post-processing experiment directory of the experiment (e.g. /some/path/pp/)

Each notebook should have a default set of model years to analyze (e.g. 1981-2010). The MAR engines will also provide two optional, additional variables, STARTYR and ENDYR, that can be used to override the defaults in the notebook.

Scalar Results / Metrics

If your notebook produces scalar metrics, it should write those results to a YAML file. See the SST_bias_NOAA_OISSTv2.ipynb notebook for an example of how to construct a YAML file. Some examples of scalar fields might be RMSE and bias of a field, or the average depth of the Mediterranean outflow plume.