Code

Code related to our research is available on GitHub, and is linked below.

Publication-related code:

• Multiyear SST prediction with CNNs: supporting code for Davenport et al. (2024). The repo includes Jupyter notebooks to train neural networks and analyze window of opportunity prediction skill. The code is written in Python, uses the Keras and TensorFlow libraries to create/train the neural network, and uses the Xarray package to analyze climate data.

• GRL2021: supporting code for Davenport and Diffenbaugh (2021). The repo includes Jupyter notebooks to recreate the analysis and figures from the paper, including training the convolutional neural network and performing trend analysis. The code is written in Python, uses the Keras and TensorFlow libraries to create/train the neural network, and uses the Xarray package to analyze climate data.

• DBD2021: code to run the fixed-effects regression models and other analysis in Davenport et al. (2021). The code is written in R.

• WRR2021: code to run the fixed-effects regression models and other analysis in Davenport et al. (2020). The code is written in R.

• Emergent constraints in CMIP6: supporting code for Simpson et al. (2021). This project started during the October 2019 CMIP6 Hackathon at NCAR, with the goal of using Python tools (like the Xarray package) to analyze climate model data. This repo includes Jupyter notebooks to recreate the analysis and figures in the paper.