Code and data for "Weather and climate forecasting with neural networks: using GCMs with different complexity as study-ground"

Code and data for the paper "Weather and climate forecasting with neural networks: using GCMs with different complexity as study-ground" in GMDD  by Sebastian Scher and Gabriele Messori

code.zip contains all code developed for the study
code/preprocessing/ contains scripts to preprocess the model data
code/training_and_prediction/ contains scripts for training and predicting with the neural networks
code/analysis contains scripts for analyzing and plotting the neural network predictions
code/clim_sims contains scripts to train the climate networks, and to produce and analyze climate runs with the networks
code/complexity contains scripts to compute and plot the local dimension

model_config_files contains configuration files for PUMA and PLASIM

data contains the data underlying the plots, and the trained networks