DiffObs: Generative Diffusion for Global Forecasting of Satellite Observations
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866909164901498880 |
|---|---|
| author | Stock, Jason Pathak, Jaideep Cohen, Yair Pritchard, Mike Garg, Piyush Durran, Dale Mardani, Morteza Brenowitz, Noah |
| author_facet | Stock, Jason Pathak, Jaideep Cohen, Yair Pritchard, Mike Garg, Piyush Durran, Dale Mardani, Morteza Brenowitz, Noah |
| contents | This work presents an autoregressive generative diffusion model (DiffObs) to predict the global evolution of daily precipitation, trained on a satellite observational product, and assessed with domain-specific diagnostics. The model is trained to probabilistically forecast day-ahead precipitation. Nonetheless, it is stable for multi-month rollouts, which reveal a qualitatively realistic superposition of convectively coupled wave modes in the tropics. Cross-spectral analysis confirms successful generation of low frequency variations associated with the Madden--Julian oscillation, which regulates most subseasonal to seasonal predictability in the observed atmosphere, and convectively coupled moist Kelvin waves with approximately correct dispersion relationships. Despite secondary issues and biases, the results affirm the potential for a next generation of global diffusion models trained on increasingly sparse, and increasingly direct and differentiated observations of the world, for practical applications in subseasonal and climate prediction. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_06517 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | DiffObs: Generative Diffusion for Global Forecasting of Satellite Observations Stock, Jason Pathak, Jaideep Cohen, Yair Pritchard, Mike Garg, Piyush Durran, Dale Mardani, Morteza Brenowitz, Noah Computational Physics Computer Vision and Pattern Recognition Machine Learning Atmospheric and Oceanic Physics This work presents an autoregressive generative diffusion model (DiffObs) to predict the global evolution of daily precipitation, trained on a satellite observational product, and assessed with domain-specific diagnostics. The model is trained to probabilistically forecast day-ahead precipitation. Nonetheless, it is stable for multi-month rollouts, which reveal a qualitatively realistic superposition of convectively coupled wave modes in the tropics. Cross-spectral analysis confirms successful generation of low frequency variations associated with the Madden--Julian oscillation, which regulates most subseasonal to seasonal predictability in the observed atmosphere, and convectively coupled moist Kelvin waves with approximately correct dispersion relationships. Despite secondary issues and biases, the results affirm the potential for a next generation of global diffusion models trained on increasingly sparse, and increasingly direct and differentiated observations of the world, for practical applications in subseasonal and climate prediction. |
| title | DiffObs: Generative Diffusion for Global Forecasting of Satellite Observations |
| topic | Computational Physics Computer Vision and Pattern Recognition Machine Learning Atmospheric and Oceanic Physics |
| url | https://arxiv.org/abs/2404.06517 |