2.2.3. PM2.5 hindcast database
The PM2.5 hindcast database (PHD) provides the PM2.5 estimation across China from 2000 to 2017. By using a machine learning approach, the PHD assembles datasets from multiple sources, including the moderate-resolution imaging spectroradiometer satellite measurements of aerosol, the Community Multiscale Air Quality model outputs based on the historical multi-resolution emission inventory for China and many other variables. It hindcasts the daily PM2.5 concentrations from 2000 to 2017 in China. The PHD is developed and maintained by a team from Tsinghua University, Beijing, China (Xue, Zheng et al. 2019).