The dataset production uses three types of data, including ground observation data, satellite remote sensing data, and reanalysis data. Among them, the ground observation data includes daily ground snow depth and snow pressure observation data from national meteorological observation stations from 2013 to 2020, as well as the National Glacier, Frozen Soil and Desert Science Data Center（ http://www.ncdc.ac.cn/ ）The ground snow depth and snow pressure observation data provided by the national snow observation stations and ground surveys from 2017 to 2019; Satellite remote sensing data includes the National Glacier, Frozen Soil and Desert Science Data Center（ http://www.ncdc.ac.cn/ ）The provided snow albedo, snow cover area, as well as SRTM digital elevation model and MODIS land vegetation classification product (MCD12Q1); Reanalysis data for ERA-5 Land dataset（ https://cds.climate.copernicus.eu ）Including meteorological factors such as wind speed and temperature, vegetation factors such as vegetation leaf area index, and snow accumulation factors such as snowfall and snowmelt.

 The dataset uses a spatiotemporal weighted neural network model to produce daily snow cover density. The input variables of the model include snow cover factors, meteorological factors, terrain factors, vegetation factors, as well as ground observation truth values. The output variable is daily snow cover density. Firstly, it is necessary to extract the snow density of observation stations and the influencing factors at their corresponding locations as samples, and use the ten fold cross validation method to determine the optimal model, and then estimate the daily snow density of the country from 2013 to 2020.

 The dataset was subjected to ten fold cross validation, and the overall R2, MAE, and RMSE of the model were 0.531, 0.028, and 0.043g/cm3, respectively. (Note: Due to the small sample size of ground stations in 2019-2020, the snow density data produced by the model is not as effective as the data from 2013-2018.)