GANterpolate: Improving Climate Modeling through Synthetic Data Generation

Keyphrases

• Synthetic Data Generation 100%
• Deep Convolutional Neural Network (deep CNN) 25%
• Reconstructed Data 25%
• Significant Region 25%
• Artificial Data 25%
• Climate Dynamics 25%
• Spatial-temporal Dependencies 25%
• Data Sparsity 25%
• Traditional Interpolation 25%
• Spatial Coherence 25%
• ERA5 Reanalysis 25%
• Equatorial Pacific 25%
• Adversarial Loss 25%

Engineering

• Synthetic Data Generation 100%
• Root Mean Square Error 100%
• Reanalysis 50%
• Sparsity 50%

Earth and Planetary Sciences

• Generative Adversarial Networks 75%
• Discriminator 25%
• Fifth Generation of ECMWF Atmospheric Reanalysis of the Global Climate 25%