A global land-based gridded dataset of daily precipitation at 1 degree X 1 degree resolution from 1950 to 2013. Two datasets are available under the REGEN moniker. This dataset interpolates only the long-term precipitation stations (stations with at least 40 complete years of data). A second related dataset interpolates all stations available regardless of completeness of station timeseries. Besides the grid cell average precipitation amount per day (mm/day), the Yamamoto standard deviation per grid cell (mm/day), the kriging error per grid cell (%) and number of stations per grid cell variables are also included.
Currently, there are two major data archives of global in situ daily rainfall data: 1. The Global Historical Station Network (GHCN-Daily) hosted by National Oceanic and Atmospheric Administration (NOAA) and 2. The Deutscher Wetterdienst (DWD) Global Precipitation Climatology Centre (GPCC). These two data archives are combined along with additional station data acquired from other researchers. The merged archive is quality controlled and the flagged stations are removed to create a long term high quality archive of raw station data. This dataset uses ordinary block kriging to interpolate a long-term station subset of this high quality merged archive. The REGEN all station based dataset instead interpolates all stations in this high quality merged archive.
The output consists of a separate CF-compliant netcdf file for each year, each containing the four aforementioned variables. These values are available for global land areas with the exception of Antarctica. The time dimension of the netcdf ranges from 1950-01-01 to 2013-12-31. Each dataset (All station based and long-term station based) is in classic netcdf format and occupies around 350MB of disk space for each year with the combined total of all years being around 26GB. Besides these two netcdfs, two additional netcdfs containing a mask indicating the high data quality grid cells of each dataset (all station and long-term) are also available. These netcdf files have the same grid descriptions as the original data but contain only one timestep for the entire period.