Scholarlink

This paper introduces the AirQ-ResUNet model, a deep learning architecture designed to emulate the EPISODE urban air quality dispersion model. The model enables accurate and efficient predictions of PM2.5 spatial distributions in urban environments, using Oslo, Norway, as a case study. AirQ-ResUNet combines a U-Net-style encoder-decoder framework with residual learning, capturing both coarse- and fine-scale patterns of pollutant dispersion. It bridges the gap between computationally intensive chemical-transport models and the demand for high-resolution air quality data in health and policy planning. By leveraging meteorological and emissions data, the model achieves a balance of precision and efficiency, preserving pollutant structures while scaling for real-world applications. Extensive validation demonstrates its ability to emulate EPISODE outputs, offering a robust tool for urban air quality forecasting, management, and decision-making. AirQ-ResUNet supports evidence-based strategies to mitigate air pollution impacts, contributing to effective public health and environmental policies.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

AirQ-ResUNet: A Residual U-Net Based Deep Learning Surrogate for High-Resolution PM2.5 Prediction in Urban Environments

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Jivitesh Sharma,
  • Islen Vallejo,
  • Rune Åvar Ødegård

摘要

This paper introduces the AirQ-ResUNet model, a deep learning architecture designed to emulate the EPISODE urban air quality dispersion model. The model enables accurate and efficient predictions of PM2.5 spatial distributions in urban environments, using Oslo, Norway, as a case study. AirQ-ResUNet combines a U-Net-style encoder-decoder framework with residual learning, capturing both coarse- and fine-scale patterns of pollutant dispersion. It bridges the gap between computationally intensive chemical-transport models and the demand for high-resolution air quality data in health and policy planning. By leveraging meteorological and emissions data, the model achieves a balance of precision and efficiency, preserving pollutant structures while scaling for real-world applications. Extensive validation demonstrates its ability to emulate EPISODE outputs, offering a robust tool for urban air quality forecasting, management, and decision-making. AirQ-ResUNet supports evidence-based strategies to mitigate air pollution impacts, contributing to effective public health and environmental policies.