GJR_hist_climate_data

This is data and code used in the paper 'How well does CMIP6 capture the dynamics of Euro-Atlantic weather regimes, and why?' by Josh Dorrington, Kristian Strommen and Federico Fabiano.

This repo contains:

Code

• code/regime_computations_example.ipynb is a Jupyter notebook demonstrating how to compute classical and geopotential-jet regimes, how to easily cluster subsets of a time series, and how to compute the stability metric. 'ClusterBuster' is a set of Python code with an accompanying example notebook showing how to compute geopotential jet regimes in ERA20C, and compute regime stability.

• An example of how to perform the ridge regression can be found at https://github.com/fedef17/FScripts/blob/master/kj_multireg_hist_3n_ensmean.py

Data

• regime_data/ contains regime patterns (as Z500 anomalies) and state sequences for both classical circulation regimes (CCR) and geopotential-jet regimes (GJR) for cluster numbers between K=2 and K=10. The focus of the paper is on GJR_K3, but all data is included for completeness.
• tuttecose_wcmip5.pkl - A python pickle file containing dictionaries of climate model predictors and regime metrics, used to fit the ridge regression model, and produce the scatter plots.

For clarification or technical help, or if you spot an error in the data/code, please contact joshua.dorrington@kit.edu