Scholarlink

<p>Global ozone pollution has increased recently. Key aspects of global ozone pollution and its health assessment are not yet fully understood owing to the limited temporal resolution of existing datasets. In this study, a global surface ozone dataset was developed, providing 0.1° daily maximum 8-hour average surface ozone for 2013–2022. An efficient Light Gradient Boosting Machine model was constructed in this study&#xa0;to combine surface ozone observations, satellite products, atmospheric chemistry model simulations, meteorological reanalysis data, emission inventories, and other geographically distributed data. Importantly, this study implemented rigorous quality control for surface ozone observations from China and Europe, while also considering temporal and spatial features of the model inputs. The dataset yielded a&#xa0;spatial-based 10-fold cross-validation coefficient of determination (R<sup>2</sup>) of 0.79–0.88 and root mean square error (RMSE) of 11.32–13.26 μg m<sup>−3</sup>, demonstrating&#xa0;an accuracy comparable to that of the current state-of-the-art dataset. Daily surface ozone datasets offer more granular information than monthly or annual datasets, thereby providing critical support for air quality management and research on ozone-related health issues.</p>

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

A global land daily 10-km-resolution surface ozone dataset from 2013–2022

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月29日

查看全文
  • Rui Wang,
  • Huanfeng Shen,
  • Chao Zeng,
  • Jiajia Chen,
  • Yuan Wang,
  • Yuyu Li

摘要

Global ozone pollution has increased recently. Key aspects of global ozone pollution and its health assessment are not yet fully understood owing to the limited temporal resolution of existing datasets. In this study, a global surface ozone dataset was developed, providing 0.1° daily maximum 8-hour average surface ozone for 2013–2022. An efficient Light Gradient Boosting Machine model was constructed in this study to combine surface ozone observations, satellite products, atmospheric chemistry model simulations, meteorological reanalysis data, emission inventories, and other geographically distributed data. Importantly, this study implemented rigorous quality control for surface ozone observations from China and Europe, while also considering temporal and spatial features of the model inputs. The dataset yielded a spatial-based 10-fold cross-validation coefficient of determination (R2) of 0.79–0.88 and root mean square error (RMSE) of 11.32–13.26 μg m−3, demonstrating an accuracy comparable to that of the current state-of-the-art dataset. Daily surface ozone datasets offer more granular information than monthly or annual datasets, thereby providing critical support for air quality management and research on ozone-related health issues.