MVAR: MultiVariate AutoRegressive Air Pollutants Forecasting Model

2025-07-16

AI Summary

Existing air pollution forecasting studies predominantly focus on single-pollutant modeling, neglecting dynamic inter-pollutant interactions and heterogeneous spatial responses, while relying on long historical input sequences—leading to low data utilization efficiency. To address these limitations, we propose MVAR, a novel framework integrating a multivariate autoregressive architecture with a meteorology-coupled spatial Transformer module. MVAR explicitly captures cross-variable pollutant dependencies and spatially heterogeneous evolution patterns, and flexibly incorporates external meteorological inputs—including ERA5 reanalysis and FuXi-2.0 weather forecasts. Crucially, it achieves 120-hour multi-step forecasting using only short temporal windows. Evaluated on a large-scale, self-constructed dataset covering 75 cities and six pollutant species, MVAR consistently outperforms state-of-the-art methods, demonstrating superior capability in modeling multivariate pollution synergies and robust long-term predictive performance.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multi-instance/Multi-view Learning
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Web Mining and Content Analysis: Web data integration and cleaning

Abstract

Air pollutants pose a significant threat to the environment and human health, thus forecasting accurate pollutant concentrations is essential for pollution warnings and policy-making. Existing studies predominantly focus on single-pollutant forecasting, neglecting the interactions among different pollutants and their diverse spatial responses. To address the practical needs of forecasting multivariate air pollutants, we propose MultiVariate AutoRegressive air pollutants forecasting model (MVAR), which reduces the dependency on long-time-window inputs and boosts the data utilization efficiency. We also design the Multivariate Autoregressive Training Paradigm, enabling MVAR to achieve 120-hour long-term sequential forecasting. Additionally, MVAR develops Meteorological Coupled Spatial Transformer block, enabling the flexible coupling of AI-based meteorological forecasts while learning the interactions among pollutants and their diverse spatial responses. As for the lack of standardized datasets in air pollutants forecasting, we construct a comprehensive dataset covering 6 major pollutants across 75 cities in North China from 2018 to 2023, including ERA5 reanalysis data and FuXi-2.0 forecast data. Experimental results demonstrate that the proposed model outperforms state-of-the-art methods and validate the effectiveness of the proposed architecture.

Problem

• Forecasting multivariate air pollutants interactions and spatial responses
• Reducing dependency on long-time-window inputs efficiently
• Addressing lack of standardized datasets for pollutants forecasting

Innovation

• Multivariate Autoregressive Training Paradigm for long-term forecasting
• Meteorological Coupled Spatial Transformer block for flexible coupling
• Comprehensive dataset covering 6 pollutants in 75 cities