Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
139 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
46 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

RainAI -- Precipitation Nowcasting from Satellite Data (2311.18398v1)

Published 30 Nov 2023 in cs.CV, cs.LG, and physics.ao-ph

Abstract: This paper presents a solution to the Weather4Cast 2023 competition, where the goal is to forecast high-resolution precipitation with an 8-hour lead time using lower-resolution satellite radiance images. We propose a simple, yet effective method for spatiotemporal feature learning using a 2D U-Net model, that outperforms the official 3D U-Net baseline in both performance and efficiency. We place emphasis on refining the dataset, through importance sampling and dataset preparation, and show that such techniques have a significant impact on performance. We further study an alternative cross-entropy loss function that improves performance over the standard mean squared error loss, while also enabling models to produce probabilistic outputs. Additional techniques are explored regarding the generation of predictions at different lead times, specifically through Conditioning Lead Time. Lastly, to generate high-resolution forecasts, we evaluate standard and learned upsampling methods. The code and trained parameters are available at https://github.com/rafapablos/w4c23-rainai.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (37)
  1. “Machine Learning for Precipitation Nowcasting from Radar Images” arXiv:1912.12132 [cs, stat] arXiv, 2019 URL: http://arxiv.org/abs/1912.12132
  2. “Recurrent Residual Convolutional Neural Network based on U-Net (R2U-Net) for Medical Image Segmentation” In CoRR abs/1802.06955, 2018 arXiv: http://arxiv.org/abs/1802.06955
  3. “Deep Learning for Day Forecasts from Sparse Observations” arXiv:2306.06079 [physics] arXiv, 2023 URL: http://arxiv.org/abs/2306.06079
  4. Georgy Ayzel, Maik Heistermann and Tanja Winterrath “Optical flow models as an open benchmark for radar-based precipitation nowcasting (rainymotion v0.1)” In Geoscientific Model Development 12.4, 2019, pp. 1387–1402 DOI: 10.5194/gmd-12-1387-2019
  5. Georgy Ayzel, Tobias Scheffer and Maik Heistermann “RainNet v1.0: a convolutional neural network for radar-based precipitation nowcasting” In Geoscientific Model Development 13.6, 2020, pp. 2631–2644 DOI: 10.5194/gmd-13-2631-2020
  6. “Accurate medium-range global weather forecasting with 3D neural networks” In Nature 619.7970, 2023, pp. 533–538 DOI: 10.1038/s41586-023-06185-3
  7. “Pangu-Weather: A 3D High-Resolution Model for Fast and Accurate Global Weather Forecast” arXiv:2211.02556 [physics] arXiv, 2022 URL: http://arxiv.org/abs/2211.02556
  8. “An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale” arXiv:2010.11929 [cs] arXiv, 2021 URL: http://arxiv.org/abs/2010.11929
  9. “Skillful Twelve Hour Precipitation Forecasts using Large Context Neural Networks” arXiv:2111.07470 [physics] arXiv, 2021 URL: http://arxiv.org/abs/2111.07470
  10. Jesús García Fernández and Siamak Mehrkanoon “Broad-UNet: Multi-scale feature learning for nowcasting tasks” In Neural Networks 144, 2021, pp. 419–427 DOI: 10.1016/j.neunet.2021.08.036
  11. “Earthformer: Exploring Space-Time Transformers for Earth System Forecasting” Publisher: arXiv Version Number: 2, 2022 DOI: 10.48550/ARXIV.2207.05833
  12. Ian Goodfellow, Yoshua Bengio and Aaron Courville “Deep Learning” MIT Press, 2016
  13. “Generative Adversarial Nets” In Advances in Neural Information Processing Systems 27 Curran Associates, Inc., 2014 URL: https://proceedings.neurips.cc/paper_files/paper/2014/file/5ca3e9b122f61f8f06494c97b1afccf3-Paper.pdf
  14. Gabriel Gouvine “NinaSR: Efficient Small and Large ConvNets for Super-Resolution” Publication Title: GitHub repository GitHub, 2021 DOI: 10.5281/zenodo.4868308
  15. “Weather4cast at NeurIPS 2022: Super-Resolution Rain Movie Prediction under Spatio-temporal Shifts”, 2022
  16. “Fully Dense UNet for 2D Sparse Photoacoustic Tomography Artifact Removal” In CoRR abs/1808.10848, 2018 arXiv: http://arxiv.org/abs/1808.10848
  17. Vincent Le Guen and Nicolas Thome “Disentangling Physical Dynamics from Unknown Factors for Unsupervised Video Prediction” arXiv:2003.01460 [cs] arXiv, 2020 URL: http://arxiv.org/abs/2003.01460
  18. “Deep Residual Learning for Image Recognition” arXiv:1512.03385 [cs] arXiv, 2015 URL: http://arxiv.org/abs/1512.03385
  19. Herman Kahn “Use of Different Monte Carlo Sampling Techniques” Santa Monica, CA: RAND Corporation, 1955
  20. “Enhanced Deep Residual Networks for Single Image Super-Resolution” arXiv:1707.02921 [cs] arXiv, 2017 URL: http://arxiv.org/abs/1707.02921
  21. “Swin Transformer: Hierarchical Vision Transformer using Shifted Windows” arXiv:2103.14030 [cs] arXiv, 2021 URL: http://arxiv.org/abs/2103.14030
  22. “Video Swin Transformer” arXiv:2106.13230 [cs] arXiv, 2021 URL: http://arxiv.org/abs/2106.13230
  23. Chuyao Luo, Xutao Li and Yunming Ye “PFST-LSTM: A SpatioTemporal LSTM Model With Pseudoflow Prediction for Precipitation Nowcasting” In IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 14, 2021, pp. 843–857 DOI: 10.1109/JSTARS.2020.3040648
  24. “Guidelines for Ensemble Prediction System” World Meteorological Organization, 2022
  25. “Attention U-Net: Learning Where to Look for the Pancreas” In CoRR abs/1804.03999, 2018 arXiv: http://arxiv.org/abs/1804.03999
  26. “WeatherFusionNet: Predicting Precipitation from Satellite Data” arXiv:2211.16824 [cs] arXiv, 2022 URL: http://arxiv.org/abs/2211.16824
  27. “Pysteps: an open-source Python library for probabilistic precipitation nowcasting (v1.0)” In Geoscientific Model Development 12.10, 2019, pp. 4185–4219 DOI: 10.5194/gmd-12-4185-2019
  28. “Skilful precipitation nowcasting using deep generative models of radar” In Nature 597.7878, 2021, pp. 672–677 DOI: 10.1038/s41586-021-03854-z
  29. Olaf Ronneberger, Philipp Fischer and Thomas Brox “U-Net: Convolutional Networks for Biomedical Image Segmentation” Publisher: arXiv Version Number: 1, 2015 DOI: 10.48550/ARXIV.1505.04597
  30. “Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting” Publisher: arXiv Version Number: 2, 2015 DOI: 10.48550/ARXIV.1506.04214
  31. “Deep Learning for Precipitation Nowcasting: A Benchmark and A New Model” arXiv:1706.03458 [cs] arXiv, 2017 URL: http://arxiv.org/abs/1706.03458
  32. “MetNet: A Neural Weather Model for Precipitation Forecasting” arXiv:2003.12140 [physics, stat] arXiv, 2020 URL: http://arxiv.org/abs/2003.12140
  33. “Ensemble Forecasting at NCEP and the Breeding Method” In Monthly Weather Review 125.12, 1997, pp. 3297–3319 DOI: 10.1175/1520-0493(1997)125<3297:EFANAT>2.0.CO;2
  34. Kevin Trebing, Tomasz Stanczyk and Siamak Mehrkanoon “SmaAt-UNet: Precipitation Nowcasting using a Small Attention-UNet Architecture” arXiv:2007.04417 [cs, eess] arXiv, 2021 URL: http://arxiv.org/abs/2007.04417
  35. “MaxViT: Multi-Axis Vision Transformer” arXiv:2204.01697 [cs] arXiv, 2022 URL: http://arxiv.org/abs/2204.01697
  36. “Attention Is All You Need” arXiv:1706.03762 [cs] arXiv, 2023 URL: http://arxiv.org/abs/1706.03762
  37. “Skilful nowcasting of extreme precipitation with NowcastNet” In Nature 619.7970, 2023, pp. 526–532 DOI: 10.1038/s41586-023-06184-4

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now
Github Logo Streamline Icon: https://streamlinehq.com

GitHub

  1. GitHub - rafapablos/w4c23-rainai: Weather4Cast 2023 NeurIPS Competition - RainAI (13 stars)