NeuroDDAF: Neural Dynamic Diffusion-Advection Fields with Evidential Fusion for Air Quality Forecasting

2026-04-01

AI Summary

This study addresses the challenges of accuracy and robustness in air pollution forecasting arising from nonlinear spatiotemporal dynamics, wind-driven transport, and regional distribution shifts. To this end, the authors propose a novel framework that integrates physical mechanisms with neural representation learning. The approach uniquely unifies open-system transport modeling with deep neural networks through a wind-modulated graph attention encoder, a Fourier-domain advection–diffusion module, a latent-space neural ordinary differential equation, and an evidential fusion mechanism. This design ensures physical consistency, cross-city generalization, and calibrated uncertainty quantification. Evaluated on four urban datasets, the model achieves a 3-day forecast RMSE as low as 48.88 μg/m³ and reduces long-term prediction errors by 9.7% in RMSE and 9.4% in MAE, substantially outperforming existing baselines.

Abstract

Accurate air quality forecasting is crucial for protecting public health and guiding environmental policy, yet it remains challenging due to nonlinear spatiotemporal dynamics, wind-driven transport, and distribution shifts across regions. Physics-based models are interpretable but computationally expensive and often rely on restrictive assumptions, whereas purely data-driven models can be accurate but may lack robustness and calibrated uncertainty. To address these limitations, we propose Neural Dynamic Diffusion-Advection Fields (NeuroDDAF), a physics-informed forecasting framework that unifies neural representation learning with open-system transport modeling. NeuroDDAF integrates (i) a GRU-Graph Attention encoder to capture temporal dynamics and wind-aware spatial interactions, (ii) a Fourier-domain diffusion-advection module with learnable residuals, (iii) a wind-modulated latent Neural ODE to model continuous-time evolution under time-varying connectivity, and (iv) an evidential fusion mechanism that adaptively combines physics-guided and neural forecasts while quantifying uncertainty. Experiments on four urban datasets (Beijing, Shenzhen, Tianjin, and Ancona) across 1-3 day horizons show that NeuroDDAF consistently outperforms strong baselines, including AirPhyNet, achieving up to 9.7% reduction in RMSE and 9.4% reduction in MAE on long-term forecasts. On the Beijing dataset, NeuroDDAF attains an RMSE of 41.63 $μ$g/m$^3$ for 1-day prediction and 48.88 $μ$g/m$^3$ for 3-day prediction, representing the best performance among all compared methods. In addition, NeuroDDAF improves cross-city generalization and yields well-calibrated uncertainty estimates, as confirmed by ensemble variance analysis and case studies under varying wind conditions.

Problem

• air quality forecasting
• spatiotemporal dynamics
• wind-driven transport
• distribution shifts
• uncertainty calibration

Innovation

• physics-informed neural networks
• diffusion-advection modeling
• Neural ODE
• evidential fusion
• spatiotemporal forecasting