Reproduce “DeepSD: Generating High Resolution Climate Change Projections through Single Image Super-Resolution” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Thomas Vandal, Evan Kodra, Sangram Ganguly, Andrew Michaelis, Ramakrishna Nemani, Auroop R Ganguly
Researchers developed DeepSD, a framework employing stacked Super-Resolution Convolutional Neural Networks, to generate high-resolution climate change projections from coarse Earth System Model outputs. This approach demonstrated improved accuracy for daily precipitation downscaling and enhanced predictability for extreme events, proving scalable for processing large climate model ensembles.
Key insights
- Super-resolution Convolutional Neural Networks can effectively learn complex, non-linear mappings between coarse and fine-resolution climate data, explicitly capturing spatial dependencies.
- Incorporating static, high-resolution topographical features as multi-scale input channels significantly enhances the predictability of downscaled climate variables.
- Framing statistical downscaling as an image super-resolution problem allows for the effective transfer of advanced deep learning architectures from computer vision to climate science.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.