A flow matching–based generative super-resolution model for global weather forecasts. Takes coarse-resolution (1.5°) forecast trajectories and generates stochastic high-resolution (0.25°) outputs, recovering fine-scale variability while preserving large-scale structure.
ArchesWeatherSR is formulated as a stochastic inverse problem using flow matching. It learns the residual between the bicubically interpolated coarse field and the true ERA5 analysis at 0.25°, concentrating model capacity on fine-scale structure. At inference, the residual is sampled and added back to the interpolated field. The backbone is a 3D Swin U-Net Transformer shared with
ArchesWeather & ArchesWeatherGen.
1git clone https://github.com/dataymeric/ArchesWeatherSR.git
2cd ArchesWeatherSR
3uv sync
4hf download dataymeric/ArchesWeatherSR --local-dir runs/archesweathersr
1from geoarches.lightning_modules import load_module
2
3sr_model, cfg = load_module("runs/archesweathersr")
4sr_model = sr_model.cuda().eval()
5
6# sample a super-resolved state from a batch
7samples = sr_model.sample(batch)
