Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Phys-Diff: A Physics-Inspired Latent Diffusion Model for Tropical Cyclone Forecasting

Lei Liu, Xiaoning Yu, Kang Chen, Jiahui Huang, Tengyuan Liu, Hongwei Zhao, Bin Li

TL;DR

Phys-Diff is proposed, a physics-inspired latent diffusion model that disentangles latent features into task-specific components and employs cross-task attention to introduce prior physics-inspired inductive biases, thereby embedding physically consistent dependencies among TC attributes.

Abstract

Tropical cyclone (TC) forecasting is critical for disaster warning and emergency response. Deep learning methods address computational challenges but often neglect physical relationships between TC attributes, resulting in predictions lacking physical consistency. To address this, we propose Phys-Diff, a physics-inspired latent diffusion model that disentangles latent features into task-specific components (trajectory, pressure, wind speed) and employs cross-task attention to introduce prior physics-inspired inductive biases, thereby embedding physically consistent dependencies among TC attributes. Phys-Diff integrates multimodal data including historical cyclone attributes, ERA5 reanalysis data, and FengWu forecast fields via a Transformer encoder-decoder architecture, further enhancing forecasting performance. Experiments demonstrate state-of-the-art performance on global and regional datasets.

Phys-Diff: A Physics-Inspired Latent Diffusion Model for Tropical Cyclone Forecasting

TL;DR

Phys-Diff is proposed, a physics-inspired latent diffusion model that disentangles latent features into task-specific components and employs cross-task attention to introduce prior physics-inspired inductive biases, thereby embedding physically consistent dependencies among TC attributes.

Abstract

Tropical cyclone (TC) forecasting is critical for disaster warning and emergency response. Deep learning methods address computational challenges but often neglect physical relationships between TC attributes, resulting in predictions lacking physical consistency. To address this, we propose Phys-Diff, a physics-inspired latent diffusion model that disentangles latent features into task-specific components (trajectory, pressure, wind speed) and employs cross-task attention to introduce prior physics-inspired inductive biases, thereby embedding physically consistent dependencies among TC attributes. Phys-Diff integrates multimodal data including historical cyclone attributes, ERA5 reanalysis data, and FengWu forecast fields via a Transformer encoder-decoder architecture, further enhancing forecasting performance. Experiments demonstrate state-of-the-art performance on global and regional datasets.
Paper Structure (19 sections, 5 equations, 4 figures, 2 tables)

This paper contains 19 sections, 5 equations, 4 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Problem Formulation
  4. Dataset
  5. Phys-Diff Architecture
  6. Conditional Denoising Network epsilon-theta
  7. PIGA Module
  8. Training Objective
  9. EXPERIMENTS
  10. Experimental Setup
  11. Dataset and Preprocessing
  12. Evaluation Metrics
  13. Implementation Details
  14. Comparison with State-of-the-Art Methods
  15. Ablation Study
  16. ...and 4 more sections

Figures (4)

  • Figure 1: Comparison of feature representations. Previous methods (e.g., MSCAR wang2024mscar, VQLTI Wang_Liu_Chen_Han_Li_Bai_2025) produce coupled features without physical constraints. Our method enforces these constraints, learning disentangled features with improved physical consistency.
  • Figure 2: Overall architecture of the Phys-Diff model. This is a denoising diffusion model built on a Transformer encoder-decoder framework. During inference, the model starts with random Gaussian noise and progressively generates predictions. The encoder handles multimodal inputs, combining historical cyclone data with environmental field features containing both past and future information. At the core, the Physics-Inspired Decoder features the PIGA module, which uses cross-task attention to model the physical dependencies between task-specific features (trajectory, wind speed, and pressure), ensuring both physical consistency and accuracy in the forecasts.
  • Figure 3: t-SNE visualization of learned task-specific features by the PIGA module.
  • Figure 4: Visualization of our trajectory forecasting results and comparison with FengWu. The five small plots visualize the ground truth (red solid line), our trajectory forecasting results (blue dashed line), and FengWu's trajectory forecasting results (green solid line), with representative trajectories selected.