ClimatExML Documentation

This documentation describes the software and training process of a statistical downscaling pipeline for climate data. It implements Super-Resolution Generative Adversarial Networks (SRGANs) following [ACM23]. The software is designed and optimized to use convection-permitting models as a high-resolution target, conditional on low-resolution inputs. The software trains Convolutional Neural Networks (CNNs) to predict high-resolution multivariate fire-weather variables for precipitation, humidity, surface temperature, and wind components.

This project is supported by the ClimatEx project.

• Preprocessing
• Installing ClimatExML
• Tracking
• Training
• Customize
• Inference
• Performance
• HPC Deployment Guide
• How to build docs

[ACM23]

Nicolaas J. Annau, Alex J. Cannon, and Adam H. Monahan. Algorithmic hallucinations of near-surface winds: statistical downscaling with generative adversarial networks to convection-permitting scales. Artificial Intelligence for the Earth Systems, 2(4):e230015, 2023. URL: https://journals.ametsoc.org/view/journals/aies/2/4/AIES-D-23-0015.1.xml, doi:https://doi.org/10.1175/AIES-D-23-0015.1.