OpenLB
Open Library of Bioscience
Advanced Search
Nature
09, 2021;597 (7878): 672-677.

Skilful precipitation nowcasting using deep generative models of radar.

Suman Ravuri, Karel Lenc, Matthew Willson, Dmitry Kangin, Remi Lam, Piotr Mirowski, Megan Fitzsimons, Maria Athanassiadou, Sheleem Kashem, Sam Madge, Rachel Prudden, Amol Mandhane, Aidan Clark, Andrew Brock, Karen Simonyan, Raia Hadsell, Niall Robinson, Ellen Clancy, Alberto Arribas, Shakir Mohamed
Author Information
  1. Suman Ravuri: DeepMind, London, UK. ORCID
  2. Karel Lenc: DeepMind, London, UK. ORCID
  3. Matthew Willson: DeepMind, London, UK. ORCID
  4. Dmitry Kangin: Met Office, Exeter, UK. ORCID
  5. Remi Lam: DeepMind, London, UK.
  6. Piotr Mirowski: DeepMind, London, UK.
  7. Megan Fitzsimons: Met Office, Exeter, UK.
  8. Maria Athanassiadou: Met Office, Exeter, UK.
  9. Sheleem Kashem: DeepMind, London, UK.
  10. Sam Madge: Met Office, Exeter, UK.
  11. Rachel Prudden: Met Office, Exeter, UK.
  12. Amol Mandhane: DeepMind, London, UK. ORCID
  13. Aidan Clark: DeepMind, London, UK.
  14. Andrew Brock: DeepMind, London, UK.
  15. Karen Simonyan: DeepMind, London, UK.
  16. Raia Hadsell: DeepMind, London, UK.
  17. Niall Robinson: Met Office, Exeter, UK.
  18. Ellen Clancy: DeepMind, London, UK.
  19. Alberto Arribas: Met Office, Exeter, UK.
  20. Shakir Mohamed: DeepMind, London, UK. shakir@deepmind.com. ORCID
PMID: 34588668 DOI: 10.1038/s41586-021-03854-z

Abstract

Precipitation nowcasting, the high-resolution forecasting of precipitation up to two hours ahead, supports the real-world socioeconomic needs of many sectors reliant on weather-dependent decision-making. State-of-the-art operational nowcasting methods typically advect precipitation fields with radar-based wind estimates, and struggle to capture important non-linear events such as convective initiations. Recently introduced deep learning methods use radar to directly predict future rain rates, free of physical constraints. While they accurately predict low-intensity rainfall, their operational utility is limited because their lack of constraints produces blurry nowcasts at longer lead times, yielding poor performance on rarer medium-to-heavy rain events. Here we present a deep generative model for the probabilistic nowcasting of precipitation from radar that addresses these challenges. Using statistical, economic and cognitive measures, we show that our method provides improved forecast quality, forecast consistency and forecast value. Our model produces realistic and spatiotemporally consistent predictions over regions up to 1,536 km × 1,280 km and with lead times from 5-90 min ahead. Using a systematic evaluation by more than 50 expert meteorologists, we show that our generative model ranked first for its accuracy and usefulness in 89% of cases against two competitive methods. When verified quantitatively, these nowcasts are skillful without resorting to blurring. We show that generative nowcasting can provide probabilistic predictions that improve forecast value and support operational utility, and at resolutions and lead times where alternative methods struggle.

References

  1. Wilson, J. W., Feng, Y., Chen, M. & Roberts, R. D. Nowcasting challenges during the Beijing Olympics: Successes, failures, and implications for future nowcasting systems. Weather Forecast. 25, 1691–1714 (2010). [DOI: 10.1175/2010WAF2222417.1]
  2. Schmid, F., Wang, Y. & Harou, A. Guidelines for Nowcasting Techniques vol. 1198 (World Meteorological Organization, 2017).
  3. Bowler, N. E., Pierce, C. E. & Seed, A. W. STEPS: a probabilistic precipitation forecasting scheme which merges an extrapolation nowcast with downscaled NWP. Quart. J. Roy. Meteorol. Soc. 132, 2127–2155 (2006). [DOI: 10.1256/qj.04.100]
  4. Pulkkinen, S. et al. PySTEPS: an open-source Python library for probabilistic precipitation nowcasting (v1. 0). Geosci. Mod. Dev. 12, 4185–4219 (2019). [DOI: 10.5194/gmd-12-4185-2019]
  5. Ayzel, G., Scheffer, T. & Heistermann, M. Rainnet v1.0: a convolutional neural network for radar-based precipitation nowcasting. Geosci. Mod. Dev. 13, 2631–2644 (2020). [DOI: 10.5194/gmd-13-2631-2020]
  6. Shi, X. et al. Deep learning for precipitation nowcasting: a benchmark and a new model. In Advances in Neural Information Processing Systems vol. 30, 5617–5627 (NeurIPS, 2017).
  7. Toth, Z. & Kalnay, E. Ensemble forecasting at NCEP and the breeding method. Mon. Weather Rev. 125, 3297–3319 (1997). [DOI: 10.1175/1520-0493(1997)125<3297]
  8. Pierce, C., Seed, A., Ballard, S., Simonin, D. & Li, Z. In Doppler Radar Observations: Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications (eds Bech, J. & Chau, J. L.) 97–142 (IntechOpen, 2012).
  9. Sun, J. Convective-scale assimilation of radar data: progress and challenges. Quart. J. Roy. Meteorol. Soc. 131, 3439–3463 (2005). [DOI: 10.1256/qj.05.149]
  10. Buehner, M. & Jacques, D. Non-Gaussian deterministic assimilation of radar-derived precipitation accumulations. Mon. Weather Rev. 148, 783–808 (2020). [DOI: 10.1175/MWR-D-19-0199.1]
  11. Harrison, D. et al. The evolution of the Met Office radar data quality control and product generation system: Radarnet. In AMS Conference on Radar Meteorology 14–18 (AMS, 2015).
  12. Germann, U. & Zawadzki, I. Scale dependence of the predictability of precipitation from continental radar images. Part II: probability forecasts. J. Appl. Meteorol. 43, 74–89 (2004). [DOI: 10.1175/1520-0450(2004)043<0074]
  13. Imhoff, R., Brauer, C., Overeem, A., Weerts, A. & Uijlenhoet, R. Spatial and temporal evaluation of radar rainfall nowcasting techniques on 1,533 events. Water Resour. Res. 56, e2019WR026723 (2020).
  14. Lebedev, V. et al. Precipitation nowcasting with satellite imagery. In International Conference on Knowledge Discovery & Data Mining 2680–2688 (ACM, 2019).
  15. Agrawal, S. et al. Machine learning for precipitation nowcasting from radar images. Preprint at https://arxiv.org/abs/1912.12132 (2019).
  16. Trebing, K., Stańczyk, T. & Mehrkanoon, S. SmaAt-UNet: precipitation nowcasting using a small attention-UNet architecture. Pattern Recog. Lett. 145, 178–186 (2021). [DOI: 10.1016/j.patrec.2021.01.036]
  17. Xingjian, S. et al. Convolutional LSTM network: a machine learning approach for precipitation nowcasting. In Advances in Neural Information Processing Systems vol. 28, 802–810 (NeurIPS, 2015).
  18. Foresti, L., Sideris, I. V., Nerini, D., Beusch, L. & Germann, U. Using a 10-year radar archive for nowcasting precipitation growth and decay: a probabilistic machine learning approach. Weather Forecast. 34, 1547–1569 (2019). [DOI: 10.1175/WAF-D-18-0206.1]
  19. Sønderby, C. K. et al. MetNet: a neural weather model for precipitation forecasting. Preprint at https://arxiv.org/abs/2003.12140 (2020).
  20. Jolliffe, I. T. & Stephenson, D. B. Forecast Verification: A Practitioner’s Guide in Atmospheric Science (John Wiley & Sons, 2012).
  21. Palmer, T. & Räisänen, J. Quantifying the risk of extreme seasonal precipitation events in a changing climate. Nature 415, 512–514 (2002). [DOI: 10.1038/415512a]
  22. Richardson, D. S. Skill and relative economic value of the ECMWF ensemble prediction system. Quart. J. Roy. Meteorol. Soc. 126, 649–667 (2000). [DOI: 10.1002/qj.49712656313]
  23. Goodfellow, I. et al. Generative adversarial nets. In Advances in Neural Information Processing Systems vol. 27, 2672–2680 (NeurIPS, 2014).
  24. Brock, A., Donahue, J. & Simonyan, K. Large scale GAN training for high fidelity natural image synthesis. In International Conference on Learning Representations (ICLR, 2019).
  25. Mirza, M. & Osindero, S. Conditional generative adversarial nets. Preprint at https://arxiv.org/abs/1411.1784 (2014).
  26. Clark, A., Donahue, J. & Simonyan, K. Adversarial video generation on complex datasets. Preprint at https://arxiv.org/abs/1907.06571 (2019).
  27. Schaefer, J. T. The critical success index as an indicator of warning skill. Weather Forecast. 5, 570–575 (1990). [DOI: 10.1175/1520-0434(1990)005<0570]
  28. Harris, D., Foufoula-Georgiou, E., Droegemeier, K. K. & Levit, J. J. Multiscale statistical properties of a high-resolution precipitation forecast. J. Hydrol. 2, 406–418 (2001).
  29. Sinclair, S. & Pegram, G. Empirical mode decomposition in 2-D space and time: a tool for space-time rainfall analysis and nowcasting. Hydrol. Earth Sys. Sci. 9, 127–137 (2005). [DOI: 10.5194/hess-9-127-2005]
  30. Gneiting, T. & Raftery, A. E. Strictly proper scoring rules, prediction, and estimation. J. Am. Stat. Assoc. 102, 359–378 (2007). [DOI: 10.1198/016214506000001437]
  31. Gilleland, E., Ahijevych, D., Brown, B. G., Casati, B. & Ebert, E. E. Intercomparison of spatial forecast verification methods. Weather Forecast. 24, 1416–1430 (2009). [DOI: 10.1175/2009WAF2222269.1]
  32. Guo, C., Pleiss, G., Sun, Y. & Weinberger, K. Q. On calibration of modern neural networks. In International Conference on Machine Learning vol. 34, 1321–1330 (ICLR, 2017).
  33. Crandall, B. W. & Hoffman, R. R. In The Oxford Handbook of Cognitive Engineering (ed. Lee, J. D.) 229–239 (Oxford Univ. Press, 2013).
  34. Fairman, J., Schultz, D., Kirshbaum, D., Gray, S. & Barrett, A. Climatology of size, shape, and intensity of precipitation features over Great Britain and Ireland. J. Hydrometeorol. 18, 1595–1615 (2017). [DOI: 10.1175/JHM-D-16-0222.1]
  35. Zhang, H., Goodfellow, I., Metaxas, D. & Odena, A. Self-attention generative adversarial networks. In International Conference on Machine Learning vol. 36, 7354–7363 (ICLR, 2019).
  36. Atger, F. The skill of ensemble prediction systems. Mon. Weather Rev. 127, 1941–1953 (1999). [DOI: 10.1175/1520-0493(1999)127<1941]
  37. Ravuri, S. & Vinyals, O. Classification accuracy score for conditional generative models. In Advances in Neural Information Processing Systems vol. 32 (NeurIPS, 2019).
  38. Kingma, D. P. & Ba, J. Adam: a method for stochastic optimization. In International Conference on Learning Representations (ICLR, 2015).
  39. Ho, J., Kalchbrenner, N., Weissenborn, D. & Salimans, T. Axial attention in multidimensional transformers. Preprint at https://arxiv.org/abs/1912.12180 (2019).
  40. Palmer, T. et al. Stochastic Parametrization and Model Uncertainty (ECMWF, 2009).
  41. Roberts, N. M. & Lean, H. W. Scale-selective verification of rainfall accumulations from high-resolution forecasts of convective events. Mon. Weather Rev. 136, 78–97 (2008). [DOI: 10.1175/2007MWR2123.1]
  42. Schwartz, C. S. et al. Toward improved convection-allowing ensembles: Model physics sensitivities and optimizing probabilistic guidance with small ensemble membership. Weather Forecast. 25, 263–280 (2010). [DOI: 10.1175/2009WAF2222267.1]
  43. Edgington, E. & Onghena, P. Randomization Tests (CRC, 2007).
  44. Bush, M. et al. The first Met Office unified model–JULES regional atmosphere and land configuration, RAL1. Geosci. Mod. Dev. 13, 1999–2029 (2020). [DOI: 10.5194/gmd-13-1999-2020]
  45. Smith, T. M. et al. Multi-radar multi-sensor (MRMS) severe weather and aviation products: Initial operating capabilities. Bull. Amer. Meteorol. Soc. 97, 1617–1630 (2016). [DOI: 10.1175/BAMS-D-14-00173.1]
  46. Jarvis, A., Reuter, H., Nelson, A. & Guevara, E. Hole-Filled Seamless SRTM Data V4 (International Centre for Tropical Agriculture, 2008).
Journal Article
Article has an altmetric score of 858

Word Cloud

Created with Highcharts 10.0.0nowcastingprecipitationmethodsgenerativeforecastoperationaldeepradarleadtimesmodelshowtwoaheadstruggleeventspredictrainconstraintsutilityproducesnowcastsprobabilisticUsingvaluepredictionsPrecipitationhigh-resolutionforecastinghourssupportsreal-worldsocioeconomicneedsmanysectorsreliantweather-dependentdecision-makingState-of-the-arttypicallyadvectfieldsradar-basedwindestimatescaptureimportantnon-linearconvectiveinitiationsRecentlyintroducedlearningusedirectlyfutureratesfreephysicalaccuratelylow-intensityrainfalllimitedlackblurrylongeryieldingpoorperformancerarermedium-to-heavypresentaddresseschallengesstatisticaleconomiccognitivemeasuresmethodprovidesimprovedqualityconsistencyrealisticspatiotemporallyconsistentregions1536 km × 1280 km5-90 minsystematicevaluation50expertmeteorologistsrankedfirstaccuracyusefulness89%casescompetitiveverifiedquantitativelyskillfulwithoutresortingblurringcanprovideimprovesupportresolutionsalternativeSkilfulusingmodels

Similar Articles

Cited By