中国高分辨率高质量PM2.5无机化学成分数据集（2013-2020）

该数据是基于开发的一种四维时空-深度森林模型，以CHAP数据集中逐日1公里无缝隙PM2.5数据集为约束，通过整合PM2.5化学成分的高密度地基观测网络，以及卫星遥感产品、气象再分析和模型模拟等数据，构建不同化学成分与总PM2.5浓度的转换模型，首次分离得到中国逐日无缝1公里无缝隙4种主要PM2.5无机化学成分卫星遥感产品，包括硫酸盐（SO42-）、硝酸盐（NO3-）、铵盐（NH4+）和氯离子（Cl-）。其中SO42-、NO3-、NH4+和Cl-日估算数据与地基观测之间的十折交叉验证决定系数R2分别为0.74，0.75，0.71和0.66,均方根误差RMSE分别为 6.0，6.6，4.3和2.3 µg/m3。主要范围为中国东部，空间分辨率为1 km，时间分辨率为日、月、年，单位为mg/m3。注意：该数据集持续更新，如需要更多数据，请发邮件联系作者（[e-mail]; [e-mail]）。

This dataset is developed based on a four-dimensional spatiotemporal deep forest model, constrained by the daily 1-km gap-free PM2.5 dataset from the CHAP dataset. By integrating high-density ground-based observation networks for PM2.5 chemical components, satellite remote sensing products, meteorological reanalysis data and model simulation data, we constructed a conversion model between different chemical components and total PM2.5 concentration. For the first time, satellite remote sensing products of four major PM2.5 inorganic chemical components at 1-km daily gap-free resolution over China were derived, including sulfate (SO₄²⁻), nitrate (NO₃⁻), ammonium (NH₄⁺), and chloride (Cl⁻). The ten-fold cross-validation determination coefficients (R²) between the daily estimated data of SO₄²⁻, NO₃⁻, NH₄⁺ and Cl⁻ and ground-based observations are 0.74, 0.75, 0.71 and 0.66 respectively, while the root mean square errors (RMSE) are 6.0, 6.6, 4.3 and 2.3 µg/m³ respectively. The dataset mainly covers eastern China, with a spatial resolution of 1 km, temporal resolutions of daily, monthly and annual, and the unit is mg/m³. Note: This dataset is continuously updated. If you need more data, please contact the authors via email ([e-mail]; [e-mail]).