Perceptually Constrained Precipitation Nowcasting Model

Wenzhi Feng, Xutao Li, Zhe Wu, Kenghong Lin, Demin Yu, Yunming Ye, Yaowei Wang
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:16777-16795, 2025.

Abstract

Most current precipitation nowcasting methods aim to capture the underlying spatiotemporal dynamics of precipitation systems by minimizing the mean square error (MSE). However, these methods often neglect effective constraints on the data distribution, leading to unsatisfactory prediction accuracy and image quality, especially for long forecast sequences. To address this limitation, we propose a precipitation nowcasting model incorporating perceptual constraints. This model reformulates precipitation nowcasting as a posterior MSE problem under such constraints. Specifically, we first obtain the posteriori mean sequences of precipitation forecasts using a precipitation estimator. Subsequently, we construct the transmission between distributions using rectified flow. To enhance the focus on distant frames, we design a frame sampling strategy that gradually increases the corresponding weights. We theoretically demonstrate the reliability of our solution, and experimental results on two publicly available radar datasets demonstrate that our model is effective and outperforms current state-of-the-art models.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-feng25h, title = {Perceptually Constrained Precipitation Nowcasting Model}, author = {Feng, Wenzhi and Li, Xutao and Wu, Zhe and Lin, Kenghong and Yu, Demin and Ye, Yunming and Wang, Yaowei}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {16777--16795}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/feng25h/feng25h.pdf}, url = {https://proceedings.mlr.press/v267/feng25h.html}, abstract = {Most current precipitation nowcasting methods aim to capture the underlying spatiotemporal dynamics of precipitation systems by minimizing the mean square error (MSE). However, these methods often neglect effective constraints on the data distribution, leading to unsatisfactory prediction accuracy and image quality, especially for long forecast sequences. To address this limitation, we propose a precipitation nowcasting model incorporating perceptual constraints. This model reformulates precipitation nowcasting as a posterior MSE problem under such constraints. Specifically, we first obtain the posteriori mean sequences of precipitation forecasts using a precipitation estimator. Subsequently, we construct the transmission between distributions using rectified flow. To enhance the focus on distant frames, we design a frame sampling strategy that gradually increases the corresponding weights. We theoretically demonstrate the reliability of our solution, and experimental results on two publicly available radar datasets demonstrate that our model is effective and outperforms current state-of-the-art models.} }
Endnote
%0 Conference Paper %T Perceptually Constrained Precipitation Nowcasting Model %A Wenzhi Feng %A Xutao Li %A Zhe Wu %A Kenghong Lin %A Demin Yu %A Yunming Ye %A Yaowei Wang %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-feng25h %I PMLR %P 16777--16795 %U https://proceedings.mlr.press/v267/feng25h.html %V 267 %X Most current precipitation nowcasting methods aim to capture the underlying spatiotemporal dynamics of precipitation systems by minimizing the mean square error (MSE). However, these methods often neglect effective constraints on the data distribution, leading to unsatisfactory prediction accuracy and image quality, especially for long forecast sequences. To address this limitation, we propose a precipitation nowcasting model incorporating perceptual constraints. This model reformulates precipitation nowcasting as a posterior MSE problem under such constraints. Specifically, we first obtain the posteriori mean sequences of precipitation forecasts using a precipitation estimator. Subsequently, we construct the transmission between distributions using rectified flow. To enhance the focus on distant frames, we design a frame sampling strategy that gradually increases the corresponding weights. We theoretically demonstrate the reliability of our solution, and experimental results on two publicly available radar datasets demonstrate that our model is effective and outperforms current state-of-the-art models.
APA
Feng, W., Li, X., Wu, Z., Lin, K., Yu, D., Ye, Y. & Wang, Y.. (2025). Perceptually Constrained Precipitation Nowcasting Model. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:16777-16795 Available from https://proceedings.mlr.press/v267/feng25h.html.

Related Material