Efficient inference and learning of a generative model for ENSO predictions from large multi-model datasets

Overview

Official Jupyter notebooks for the implementation of a variational autoencoder (VAE) for ENSO modeling and prediction.

The work is published in

Groth and Chavez, 2024: Efficient inference and learning of a generative model for ENSO predictions from large multi-model datasets. Climate Dynamics, https://doi.org/10.1007/s00382-024-07162-w.

In this paper, historical simulations of global sea-surface temperature (SST) from the fifth phase of the Coupled Model Intercomparison Project (CMIP5) are analyzed. Based on the concept of a variational auto-encoder (VAE), a generative model of global SST is proposed in combination with an inference model that aims to solve the problem of determining a joint distribution over the data generating factors. With a focus on the El Niño Southern Oscillation (ENSO), the performance of the VAE-based approach in simulating various central features of observed ENSO dynamics is demonstrated. A combination of the VAE with a forecasting model is proposed to make predictions about the distribution of global SST and the corresponding future path of the Niño index from the learned latent factors.

Requirements

1. The Jupyter notebooks require the VAE package, which is available at:

   https://github.com/andr-groth/VAE-project

2. Sample data used in Groth and Chavez (2024) is included in the data/ folder. The data was collected with the help of the Climate Explorer at:

   http://climexp.knmi.nl

   For more information on the data see data/README.md.

Usage

The repository includes two Jupyter notebooks, one for model training and another one for model exploration.

1. The model is trained with VAEp_train.ipynb
2. Properties of the trained model are explored with VAEp_explore.ipynb

The weights of a trained model used to create the figures in Groth and Chavez (2024) are provided in the logs/ folder of this repository.

Example runs of the Jupyter notebooks are available in the examples/ folder of this repository. The examples are based on the sample data in the data/ folder.

Reference

Please add a reference to the following paper if you use parts of this code:

@Article{Groth.Chavez.2024,
  author           = {Groth, Andreas and Chavez, Erik},
  journal          = {Climate Dynamics},
  title            = {Efficient inference and learning of a generative model for {ENSO} predictions from large multi-model datasets},
  year             = {2024},
  doi              = {10.1007/s00382-024-07162-w},
  publisher        = {Springer Science and Business Media LLC},
}