Product overview

产品概述

This product provides high-resolution, species-resolved ambient VOC concentrations for the Beijing–Tianjin–Hebei region. It combines ground-based VOC observations, satellite retrievals, chemical transport model simulations, meteorological reanalysis, and surface and socioeconomic information within a unified machine-learning framework. The inputs were harmonized to a common spatiotemporal grid, and the resulting estimates were evaluated through internal and external validation to produce a seamless regional dataset.

本产品面向京津冀地区，融合地面VOCs在线观测、卫星遥感、化学传输模型模拟、再分析气象数据以及地表和人口经济信息，通过统一的机器学习框架构建高时空分辨率环境空气VOCs分物种浓度数据集。各类输入数据经时空尺度统一后进行反演，并通过内部与外部验证形成区域无缝数据产品。

Product characteristics

产品特点

• Multi-source data fusion: combines ground observations, satellite retrievals, chemical transport modelling, meteorological reanalysis, and surface and socioeconomic information.
• Species-resolved representation: covers major VOC groups, including alkanes, alkenes, alkynes, aromatics, and oxygenated VOCs.
• Consistent modelling framework: applies XGBoost across the 14 machine-learning-derived species to improve methodological consistency and inter-species comparability.
• Multi-level validation: model performance was assessed using five-fold cross-validation, an independent test set, and external observations reported in the literature.
• Observation-constrained estimates: combines CMAQ-based chemical and transport information with ground observations to reduce reliance on a single data source.

• 多源数据融合：综合地面监测、卫星遥感、化学传输模型、再分析气象数据以及地表和人口经济信息。
• 分物种精细表征：覆盖烷烃、烯烃、炔烃、芳香烃和含氧挥发性有机物等主要VOCs类别。
• 统一反演框架：对14种机器学习反演物种统一采用XGBoost模型，增强不同物种结果的方法一致性和可比性。
• 多层次模型验证：通过五折交叉验证、独立测试集和文献外部观测对模型性能进行评估。
• 观测约束估计：结合CMAQ提供的化学与传输信息和地面观测约束，降低对单一数据源的依赖。

Dataset preview

数据集图片

Suggested citation

建议引用