PrecipJEPA: JEPA-Regularized Future-State Prediction with Motion-Source Rendering for Precipitation Nowcasting
Not on the map. Why?
Abstract (arXiv)
Long-term precipitation nowcasting requires modeling radar-echo evolution while preserving localized high-intensity structures. Recent radar-specific studies motivate location-aware prediction and separating echo displacement from intensity change. However existing encoders learn historical representations mainly from final forecast errors. We propose PrecipJEPA, which couples a structured forecasting path with an auxiliary path that enriches its encoder from observed radar history. In the forecasting path, an online encoder first converts the observations into spatiotemporal tokens. The Task-Driven Future-State Predictor (TFP) combines these tokens with a recent-dynamics summary and spatiotemporal queries to construct future radar states. The Parallel Motion-Source Renderer (PMSR) decodes these states into motion and source-sink fields that transform the latest observation into future frames. During joint training, the History-Masked JEPA (H-JEPA) operates on the auxiliary path to predict masked historical features from visible context, directly supervising the same online encoder from the observed sequence. Experiments on SEVIR and MeteoNet show that PrecipJEPA improves highest-threshold CSI by 118.6% and 35.1%, respectively, over the strongest baselines, while maintaining the highest mean CSI throughout the 3-hour forecast.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 14 of its 22 references that are in the map. First posted on arXiv 30 Sep 2026.
Builds on
14 of 22 references.
References · 14
- 2015Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcastingcited by 1,436
- 2023Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecturecited by 1,039
- 2024Revisiting Feature Prediction for Learning Visual Representations from Videocited by 531
- 2021Skillful Precipitation Nowcasting using Deep Generative Models of Radarcited by 311
- 2022SimVP: Simpler yet Better Video Predictioncited by 184
- 2017Deep Learning for Precipitation Nowcasting: A Benchmark and A New Modelcited by 182
- 2021PredRNN: A Recurrent Neural Network for Spatiotemporal Predictive Learningcited by 182
- 2022Earthformer: Exploring Space-Time Transformers for Earth System Forecastingcited by 152
- 2023PreDiff: Precipitation Nowcasting with Latent Diffusion Modelscited by 98
- 2023DiffCast: A Unified Framework via Residual Diffusion for Precipitation Nowcastingcited by 61
- 2024CasCast: Skillful High-resolution Precipitation Nowcasting via Cascaded Modellingcited by 56
- 2025FlowCast: Advancing Precipitation Nowcasting with Conditional Flow Matchingcited by 8
Built on by
Not cited yet.