# Precipitation nowcasting of satellite data using physically-aligned neural networks

arXiv: https://arxiv.org/abs/2511.05471 · HTML: https://mlsanity.com/paper/2511.05471/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 21 of its 48 references that are in the map.
- First posted on arXiv 7 Nov 2025.

## Builds on (most cited first, 21 of 21)

- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting](https://mlsanity.com/paper/1506.04214/) (2015)
- [When and why PINNs fail to train: A neural tangent kernel perspective](https://mlsanity.com/paper/2007.14527/) (2020)
- [MaxViT: Multi-Axis Vision Transformer](https://mlsanity.com/paper/2204.01697/) (2022)
- [Skillful Precipitation Nowcasting using Deep Generative Models of Radar](https://mlsanity.com/paper/2104.00954/) (2021)
- [Adaptive Fourier Neural Operators: Efficient Token Mixers for Transformers](https://mlsanity.com/paper/2111.13587/) (2021)
- [Self-Adaptive Physics-Informed Neural Networks using a Soft Attention Mechanism](https://mlsanity.com/paper/2009.04544/) (2020)
- [FourCastNet: Accelerating Global High-Resolution Weather Forecasting using Adaptive Fourier Neural Operators](https://mlsanity.com/paper/2208.05419/) (2022)
- [PredRNN: A Recurrent Neural Network for Spatiotemporal Predictive Learning](https://mlsanity.com/paper/2103.09504/) (2021)
- [Earthformer: Exploring Space-Time Transformers for Earth System Forecasting](https://mlsanity.com/paper/2207.05833/) (2022)
- [PreDiff: Precipitation Nowcasting with Latent Diffusion Models](https://mlsanity.com/paper/2307.10422/) (2023)
- [Deep Learning for Day Forecasts from Sparse Observations](https://mlsanity.com/paper/2306.06079/) (2023)
- [Generative Adversarial Networks (GANs Survey): Challenges, Solutions, and Future Directions](https://mlsanity.com/paper/2005.00065/) (2020)
- [CasCast: Skillful High-resolution Precipitation Nowcasting via Cascaded Modelling](https://mlsanity.com/paper/2402.04290/) (2024)
- [Conditional physics informed neural networks](https://mlsanity.com/paper/2104.02741/) (2021)
- [Precipitation Nowcasting with Satellite Imagery](https://mlsanity.com/paper/1905.09932/) (2019)
- [Deep learning for precipitation nowcasting: A survey from the perspective of time series forecasting](https://mlsanity.com/paper/2406.04867/) (2024)
- [Precipitation Nowcasting Using Physics Informed Discriminator Generative Models](https://mlsanity.com/paper/2406.10108/) (2024)
- [Data-driven Precipitation Nowcasting Using Satellite Imagery](https://mlsanity.com/paper/2412.11480/) (2024)
- [An Operational Deep Learning System for Satellite-Based High-Resolution Global Nowcasting](https://mlsanity.com/paper/2510.13050/) (2025)
- [Global Precipitation Nowcasting of Integrated Multi-satellitE Retrievals for GPM: A U-Net Convolutional LSTM Architecture](https://mlsanity.com/paper/2307.10843/) (2023)

## Built on by (most cited first, 1 of 1)

- [PixelFlowCast: Latent-Free Precipitation Nowcasting via Pixel Mean Flows](https://mlsanity.com/paper/2605.10046/) (2026)

## Abstract (arXiv)

Accurate short-term precipitation forecasts predominantly rely on dense weather-radar networks, limiting operational value in places most exposed to climate extremes. We present TUPANN (Transferable and Universal Physics-Aligned Nowcasting Network), a satellite-only model trained on GOES-16 RRQPE. Unlike most deep learning models for nowcasting, TUPANN decomposes the forecast into physically meaningful components: a variational encoder-decoder infers motion and intensity fields from recent imagery under optical-flow supervision, a lead-time-conditioned MaxViT evolves the latent state, and a differentiable advection operator reconstructs future frames. We evaluate TUPANN on both GOES-16 and IMERG data, in up to four distinct climates (Rio de Janeiro, Manaus, Miami, La Paz) at 10-180min lead times using the CSI and HSS metrics over 4-64 mm/h thresholds. Comparisons against optical-flow, deep learning and hybrid baselines show that TUPANN achieves the best or second-best skill in most settings, with pronounced gains at higher thresholds. Training on multiple cities further improves performance, while cross-city experiments show modest degradation and occasional gains for rare heavy-rain regimes. The model produces smooth, interpretable motion fields aligned with numerical optical flow and runs in near real time due to the low latency of GOES-16. These results indicate that physically aligned learning can provide nowcasts that are skillful, transferable and global.