Models and Predictions for "Diffusion-Based Probabilistic Modeling for Hourly Streamflow Prediction and Assimilation"

# Models and Predictions for "Diffusion-Based Probabilistic Modeling for Hourly Streamflow Prediction and Assimilation"This repository contains the results of the paper **"Diffusion-Based Probabilistic Modeling for Hourly Streamflow Prediction and Assimilation"** (https://doi.org/10.1029/2025WR042720).`models`:- `single_forcing_model`: the checkpoint and config of the single-forcing diffusion model- `multi_forcing_model`: the checkpoint and config of the multiple-forcing diffusion model- `inverse_model`: the checkpoint and config of the streamflow-to-precipitation inverse diffusion model`predictions`:- `pred_single_forcing`: ensemble streamflow predictions of the single-forcing diffusion model- `pred_siungle_forcing_inpainting`: ensemble streamflow predictions of the single-forcing diffusion model with inpainting for training‐free data assimilation- `pred_multi_forcing`: ensemble streamflow predictions of the multiple-forcing diffusion model- `pred_multi_forcing_inpainting`: ensemble streamflow predictions of the multiple-forcing diffusion model with inpainting for training‐free data assimilation- `precip_inverse_generation`: ensemble precipitation predictions of the inverse diffusion model with inpainting for flexible conditional generation`daily_sim`:Daily streamflow simulations using the dhbv model (https://zenodo.org/records/17055819)- `dhbv_1990-2014_ep50_daymet.nc`: dhbv streamflow simulations using daymet forcing- `dhbv_1990-2014_ep50_nldas.nc`: dhbv streamflow simulations using nldas forcing- `dhbv_1990-2008_ep50_multiforcing`: dhbv streamflow simulations using daymet/nldas/maurer forcing separately