本研究采用基于随机森林的回溯模拟方法，生成了 2000-2020 年中国区域逐日 1 km全覆盖 PM2.5数据集。2013 年以前国内缺乏可用的 PM2.5地面观测数据，无法直接用于模型构建与精度验证，因此本方法重点优化了该时段的 PM2.5估算效果。研究首次将 2013 年之前的观测预测因子纳入模型运算，模型输入数据涵盖多类数据源：MAIAC 气溶胶光学厚度、中国气象局气象观测数据、ERA-5 再分析数据及其他陆面相关数据。\n
本数据集为2000-2022年月均数据，格式为 GEOTIFF；2021-2022年数据沿用上述文献所构建的模型独立估算完成。经样本十折交叉验证，2021 年模型决定系数R2为 0.91，均方根误差 RMSE 为 8.84 μg/m3；2022 年R2为 0.93，RMSE 为 7.42 μg/m3。", "ds_source": "", "ds_process_way": "", "ds_quality": "", "ds_acq_start_time": "2000-01-01 00:00:00", "ds_acq_end_time": "2022-12-31 00:00:00", "ds_acq_place": "中国", "ds_acq_lon_east": null, "ds_acq_lat_south": null, "ds_acq_lon_west": null, "ds_acq_lat_north": null, "ds_acq_alt_low": null, "ds_acq_alt_high": null, "ds_share_type": "open-access", "ds_total_size": 8727859066, "ds_files_count": 0, "ds_format": "GeoTIFF", "ds_space_res": "1km", "ds_time_res": "日", "ds_coordinate": "无", "ds_projection": "", "ds_thumbnail": "54dbfe3d-56a6-41e9-8de6-321dca692329.png", "ds_thumb_from": 2, "ds_ref_way": "", "paper_ref_way": "", "ds_ref_instruction": "", "ds_from_station": null, "organization_id": "0a4269e1-65f4-45f1-aeba-88ea3068eebf", "ds_serv_man": "李红星", "ds_serv_phone": "0931-4967592", "ds_serv_mail": "ncdc@lzb.ac.cn", "doi_value": "", "subject_codes": [ "170.15" ], "quality_level": 0, "publish_time": "2026-05-28 11:12:34", "last_updated": "2026-05-28 11:12:34", "protected": false, "protected_to": null, "lang": "zh", "cstr": "11738.11.ncdc.atmosphere.db7401.2026", "i18n": { "en": { "title": "Gap-free 1-km PM2.5 dataset in China (2000-2022)", "ds_format": "GeoTIFF", "ds_source": "", "ds_quality": "", "ds_ref_way": "", "ds_abstract": "
This study uses a random forest-based retrospective simulation method to generate a PM2.5 dataset with daily full coverage of 1 kilometer in China from 2000 to 2020.Before 2013, there was a lack of available PM2.5 ground observation data in China and could not be directly used for model construction and accuracy verification. Therefore, this method focuses on optimizing the PM2.5 estimation effect during this period.For the first time, the study incorporated observed prediction factors before 2013 into the model operation. The model input data covers multiple types of data sources: MAIAC aerosol optical thickness, meteorological observation data from China Meteorological Administration, ERA-5 reanalysis data and other land-related data.\r\n
This dataset is monthly average data from 2000 to 2022 in the format GEOTIFF; the data from 2021 to 2022 are independently estimated using the model constructed in the above documents. After cross-verification with ten folds of samples, the model's coefficient of determination R2 in 2021 is 0.91, and the root-mean-square error RMSE is 8.84 μg/m3; in 2022, the R2 is 0.93, and the RMSE is 7.42 μg/m3.", "ds_time_res": "", "ds_acq_place": "China", "ds_space_res": "", "ds_projection": "", "ds_process_way": "", "ds_ref_instruction": "" } }, "submit_center_id": "ncdc", "data_level": 0, "recommendation_value": 0, "license_type": "https://creativecommons.org/licenses/by/4.0/", "doi_reg_from": "reg_local", "cstr_reg_from": "reg_local", "doi_not_reg_reason": null, "cstr_not_reg_reason": null, "is_paper_in_submitting": false, "belong_to_nieer": false, "ds_topic_tags": [ "PM2.5", "逐日", "长序列" ], "ds_subject_tags": [ "大气科学" ], "ds_class_tags": [], "ds_locus_tags": [ "中国" ], "ds_time_tags": [ 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012, 2013, 2014, 2015, 2016, 2017, 2018, 2019, 2020, 2021, 2022 ], "ds_contributors": [ { "true_name": "何青青", "email": "qqhe@whut.edu.cn", "work_for": "武汉理工大学资源与环境工程学院", "country": "中国" } ], "ds_meta_authors": [ { "true_name": "何青青", "email": "qqhe@whut.edu.cn", "work_for": "武汉理工大学资源与环境工程学院", "country": "中国" } ], "ds_managers": [ { "true_name": "何青青", "email": "qqhe@whut.edu.cn", "work_for": "武汉理工大学资源与环境工程学院", "country": "中国" } ], "category": "大气本底" }