# Precipitation Forecasting

A separately trained diffusion generator maps non-precipitation atmospheric states from a frozen weather model to observation-aligned 24-hour precipitation, improving daily accumulation and event detection on Pangu-Weather, GraphCast and HRES without backbone-specific retraining.

A separately trained diffusion generator maps non-precipitation atmospheric states from a frozen weather model to observation-aligned 24-hour precipitation, improving daily accumulation and event detection on Pangu-Weather, GraphCast and HRES without backbone-specific retraining.

The authors introduce SoftSEEPS, a sigmoid relaxation of the SEEPS dry/light/heavy categorical score that is differentiable and usable directly as a training loss; training a convolutional decoder on the latent space of a frozen ArchesWeather-M backbone for IMERG daily precipitation, validation discrete SEEPS falls from 0.481 with MSE only to 0.375 with SoftSEEPS only, and to 0.381 under a joint objective while RMSE rises only from 5.316 to 5.334.

The authors introduce SoftSEEPS, a sigmoid relaxation of the SEEPS dry/light/heavy categorical score that is differentiable and usable directly as a training loss; training a convolutional decoder on the latent space of a frozen ArchesWeather-M backbone for IMERG daily precipitation, validation discrete SEEPS falls from 0.481 with MSE only to 0.375 with SoftSEEPS only, and to 0.381 under a joint objective while RMSE rises only from 5.316 to 5.334.